CN117813846A - Apparatus, method and computer program - Google Patents

Abstract

An apparatus comprising means configured to: facilitating cross-domain network service related machine learning or artificial intelligence pipeline reliability functionality, wherein the cross-domain machine learning or artificial intelligence pipeline is used to control a cognitive autonomous network comprising at least two domains.

Description

Device, method and computer program

Technical Field

The present disclosure relates to apparatus, methods, and computer programs for providing a framework for cross-domain trusted artificial intelligence applications applicable to, but not exclusively, cognitive autonomous networks.

Background Art

A communication system may be viewed as a facility that enables communication sessions between two or more entities (such as communication devices, base stations, and/or other nodes) by providing carrier waves between the various entities involved in the communication path.

The communication system may be a wireless communication system. Examples of wireless systems include public land mobile networks (PLMNs) operating based on radio standards such as those provided by 3GPP, satellite-based communication systems, and different wireless local area networks (e.g., wireless local area networks (WLANs)). Wireless systems can typically be divided into cells and are therefore often referred to as cellular systems.

Communication systems and associated devices typically operate according to a given standard or specification that specifies what the various entities associated with the system are allowed to do and how it should be implemented. The communication protocols and/or parameters that should be used for the connection are also typically defined. An example of a standard is the so-called 5G standard.

There is a need to provide a control system that enables a communication service provider (CSP) to control and optimize a complex network of communication system elements.

One of the current approaches being adopted is closed-loop automation and machine learning, which can be built into self-organizing networks (SONs), enabling operators to automatically optimize each cell in a radio access network.

Summary of the invention

According to a first aspect, an apparatus is provided, comprising components configured to facilitate machine learning or artificial intelligence pipeline credibility functions related to cross-domain network services, wherein the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network including at least two domains.

A component configured to facilitate cross-domain network service-related machine learning or artificial intelligence pipeline credibility functionality may be configured to facilitate at least one of the following, wherein the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network comprising at least two domains: defining cross-domain network service-related machine learning or artificial intelligence pipeline credibility; configuring cross-domain network service-related machine learning or artificial intelligence pipeline credibility; measuring cross-domain network service-related machine learning or artificial intelligence pipeline credibility; and reporting machine learning or artificial intelligence pipeline credibility associated with cross-domain network services.

The machine learning or artificial intelligence pipeline credibility function associated with cross-domain network services may include at least one of the following: fairness; explainability; and robustness.

Components configured to facilitate cross-domain network service-related machine learning or artificial intelligence pipeline credibility functions can be configured for the following items, where the cross-domain machine learning or artificial intelligence pipeline is used for controlling a cognitive autonomous network including at least two domains: obtaining cross-domain machine learning or artificial intelligence quality credibility, the cross-domain machine learning or artificial intelligence quality credibility is configured to: cover domain-specific machine learning or artificial intelligence quality credibility requirements, and constraints of the cross-domain network service-related machine learning or artificial intelligence pipeline; convert the cross-domain machine learning or artificial intelligence quality credibility into at least one domain-specific machine learning or artificial intelligence quality credibility for at least one domain of the at least two domains; and provide at least one domain-specific machine learning or artificial intelligence quality credibility for at least one domain of the at least two domains.

A component configured to convert cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality for at least one of at least two domains can be configured to convert the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality based on the risk level of the cross-domain network service.

A component configured to convert cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality for at least one of at least two domains can be configured to convert the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality based on at least one service level agreement requirement for a cross-domain network, wherein at least one service level agreement includes at least one of the following: service type; service priority; and at least one key performance indicator metric.

A component configured to provide at least one domain-specific machine learning or artificial intelligence credibility quality for at least one domain of at least two domains can be configured to: generate a cross-domain credibility machine learning or artificial intelligence configuration or delegation request and pass it to at least one domain-specific machine learning or artificial intelligence credibility function, the cross-domain credibility machine learning or artificial intelligence configuration or delegation request being configured to control at least one domain-specific machine learning or artificial intelligence credibility function to configure a machine learning or artificial intelligence pipeline for at least one domain of at least two domains; and obtain a cross-domain credibility machine learning or artificial intelligence configuration or delegation response from at least one domain-specific machine learning or artificial intelligence credibility function based on the implementation of the configuration of the machine learning or artificial intelligence pipeline for at least one domain of at least two domains.

A component configured to provide at least one domain-specific machine learning or artificial intelligence credibility quality for at least one domain of at least two domains can be configured to: generate a cross-domain credibility machine learning or artificial intelligence configuration or delegation request and pass it to at least one domain-specific policy manager, the cross-domain credibility machine learning or artificial intelligence configuration or delegation request being configured to control at least one domain-specific machine learning or artificial intelligence credibility function to configure a machine learning or artificial intelligence pipeline for at least one domain of at least two domains; and obtain a cross-domain credibility machine learning or artificial intelligence configuration or delegation response from at least one domain policy manager based on the implementation of the configuration of the machine learning or artificial intelligence pipeline for at least one domain of at least two domains.

A cross-domain trustworthy machine learning or artificial intelligence configuration or delegation request may include: a domain scope parameter configured to identify the domain to which the request is being addressed; a pipeline identification parameter configured to identify the domain-specific machine learning or artificial intelligence pipeline to which the request is being addressed; and a category identification parameter configured to identify the at least one domain-specific machine learning or artificial intelligence trustworthiness quality.

The cross-domain trustworthy machine learning or artificial intelligence configuration or delegation request may also include at least one of the following: an expected fairness parameter configured to indicate a relative fairness level for a domain-specific machine learning or artificial intelligence pipeline; an expected explainability parameter configured to indicate an expected explainability level for a domain-specific machine learning or artificial intelligence pipeline; an expected technical robustness parameter configured to indicate an expected technical robustness level for a domain-specific machine learning or artificial intelligence pipeline; and an expected adversarial robustness parameter configured to indicate an expected adversarial robustness level for a domain-specific machine learning or artificial intelligence pipeline.

A component configured to facilitate machine learning or artificial intelligence pipeline trustworthiness functions related to cross-domain network services can be configured for the following items, where the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network including at least two domains: generating a cross-domain trustworthiness machine learning or artificial intelligence capability information request and passing it to at least one domain-specific machine learning or artificial intelligence trustworthiness function, wherein the cross-domain trustworthiness machine learning or artificial intelligence capability information request is configured to control at least one domain-specific machine learning or artificial intelligence trustworthiness function to enable capability discovery of a machine learning or artificial intelligence pipeline for at least one of the at least two domains; and obtaining a cross-domain trustworthiness machine learning or artificial intelligence capability information response from at least one domain-specific machine learning or artificial intelligence trustworthiness function, reporting capability discovery for a machine learning or artificial intelligence pipeline for at least one of the at least two domains.

A cross-domain trustworthy machine learning or artificial intelligence capability information request may include: a domain scope parameter configured to identify the domain to which the request is being addressed; and a scope parameter configured to identify a domain-specific machine learning or artificial intelligence pipeline to which the request is being addressed.

The cross-domain trustworthiness machine learning or artificial intelligence capability information request may also include a pipeline stage parameter, which is configured to identify the stage of the domain-specific machine learning or artificial intelligence pipeline that the request is addressing.

Components configured to facilitate machine learning or artificial intelligence pipeline credibility functions associated with cross-domain network services may also be configured for the following items, wherein the cross-domain machine learning or artificial intelligence pipeline is used to control a cognitive autonomous network comprising at least two domains: obtaining a cross-domain credibility artificial intelligence report request or subscription from a network operator; generating a domain-specific credibility artificial intelligence report request or subscription based on the cross-domain credibility artificial intelligence report request or subscription from the network operator and passing it to at least one domain-specific machine learning or artificial intelligence pipeline credibility function, wherein the domain-specific credibility artificial intelligence report request or subscription is configured to control at least one domain-specific machine learning or artificial intelligence pipeline credibility function to provide at least one domain-specific machine learning or artificial intelligence pipeline report response or notification; receiving machine learning or artificial intelligence capability information and/or at least one domain-specific machine learning or artificial intelligence pipeline report response or notification from at least one domain-specific machine learning or artificial intelligence pipeline credibility function; storing the machine learning or artificial intelligence capability information and/or at least one domain-specific machine learning or artificial intelligence pipeline report received from at least one domain-specific machine learning or artificial intelligence pipeline in a cross-domain trust knowledge database; generating and passing a cross-domain credibility artificial intelligence report based on at least one domain-specific machine learning or artificial intelligence pipeline report response or notification.

A cross-domain credibility AI report request may include: a domain scope parameter configured to identify the domain to which the request is being addressed; a scope parameter configured to identify a domain-specific machine learning or AI pipeline to which the request is being addressed; and a pipeline stage parameter configured to identify a stage of a domain-specific machine learning or AI pipeline to which the request is being addressed. The cross-domain trustworthy AI report request may also include at least one of the following: a list of fairness metric parameters, configured to identify fairness metrics to be reported; a list of fairness metric explanations, configured to identify fairness metric explanations to be reported; a list of explainability metrics, configured to identify explainability metrics to be reported; a list of explanations, configured to identify explanations to be reported; a list of technical robustness metrics, configured to identify technical robustness metrics to be reported; a list of technical robustness metric explanations, configured to identify technical robustness metric explanations to be reported; a list of adversarial robustness metrics, configured to identify adversarial robustness metrics to be reported; a list of adversarial robustness metric explanations, configured to identify adversarial robustness metric explanations to be reported; a start time parameter, configured to identify a start time for reporting; an end time parameter, configured to identify an end time for reporting; and a reporting interval parameter, configured to identify a periodic interval for reporting.

A cross-domain credibility AI report subscription may include: a domain scope parameter configured to identify the domain that the subscription is addressing; a scope parameter configured to identify the domain-specific machine learning or AI pipeline that the subscription is addressing; and a pipeline stage parameter configured to identify the stage of the domain-specific machine learning or AI pipeline that the subscription is addressing.

The cross-domain trustworthy AI report subscription may also include at least one of the following: a list of fairness metric parameters, configured to identify fairness metrics to be reported; a list of explainability metrics, configured to identify explainability metrics to be reported; a list of technical robustness metrics, configured to identify technical robustness metrics to be reported; a list of adversarial robustness metrics, configured to identify adversarial robustness metrics to be reported; and a cross-reporting threshold parameter, configured to identify a threshold against which a metric or metric interpretation is reported.

According to a second aspect, a method is provided, comprising: facilitating machine learning or artificial intelligence pipeline credibility functions related to cross-domain network services, wherein the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network including at least two domains.

Facilitating cross-domain network service-related machine learning or artificial intelligence pipeline trustworthiness functionality may include facilitating at least one of the following, where the cross-domain machine learning or artificial intelligence pipeline is used to control a cognitive autonomous network that includes at least two domains: defining cross-domain network service-related machine learning or artificial intelligence pipeline trustworthiness; configuring cross-domain network service-related machine learning or artificial intelligence pipeline trustworthiness; measuring cross-domain network service-related machine learning or artificial intelligence pipeline trustworthiness; and reporting machine learning or artificial intelligence pipeline trustworthiness associated with cross-domain network services.

The machine learning or artificial intelligence pipeline credibility function associated with cross-domain network services may include at least one of the following: fairness; explainability; and robustness.

Functionality for facilitating cross-domain network service-related machine learning or artificial intelligence pipeline credibility may include, where the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network comprising at least two domains: obtaining cross-domain machine learning or artificial intelligence quality credibility, the cross-domain machine learning or artificial intelligence quality credibility being configured to: cover domain-specific machine learning or artificial intelligence quality credibility requirements and constraints of the cross-domain network service-related machine learning or artificial intelligence pipeline; converting the cross-domain machine learning or artificial intelligence quality credibility into at least one domain-specific machine learning or artificial intelligence quality credibility for at least one domain of the at least two domains; and providing at least one domain-specific machine learning or artificial intelligence quality credibility for at least one domain of the at least two domains.

Converting the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality for at least one domain of at least two domains may include converting the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality based on the risk level of the cross-domain network service.

Converting the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality for at least one domain of at least two domains may include converting the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality based on at least one service level agreement requirement for the cross-domain network, wherein at least one service level agreement includes at least one of the following: service type; service priority; and at least one key performance indicator metric.

Providing at least one domain-specific machine learning or artificial intelligence credibility quality for at least one domain of at least two domains may include: generating a cross-domain credibility machine learning or artificial intelligence configuration or delegation request and passing it to at least one domain-specific machine learning or artificial intelligence credibility function, the cross-domain credibility machine learning or artificial intelligence configuration or delegation request being configured to control at least one domain-specific machine learning or artificial intelligence credibility function to configure a machine learning or artificial intelligence pipeline for at least one domain of at least two domains; and obtaining a cross-domain credibility machine learning or artificial intelligence configuration or delegation response from at least one domain-specific machine learning or artificial intelligence credibility function based on the implementation of the configuration of the machine learning or artificial intelligence pipeline for at least one domain of at least two domains.

Providing at least one domain-specific machine learning or artificial intelligence credibility quality for at least one domain of at least two domains may include: generating a cross-domain credibility machine learning or artificial intelligence configuration or delegation request and passing it to at least one domain-specific policy manager, the cross-domain credibility machine learning or artificial intelligence configuration or delegation request being configured to control at least one domain-specific machine learning or artificial intelligence credibility function to configure a machine learning or artificial intelligence pipeline for at least one domain of at least two domains; and obtaining a cross-domain credibility machine learning or artificial intelligence configuration or delegation response from at least one domain policy manager based on the implementation of the configuration of the machine learning or artificial intelligence pipeline for at least one domain of at least two domains.

A cross-domain trustworthy machine learning or artificial intelligence configuration or delegation request may include: a domain scope parameter configured to identify the domain to which the request is being addressed; a pipeline identification parameter configured to identify the domain-specific machine learning or artificial intelligence pipeline to which the request is being addressed; and a category identification parameter configured to identify the at least one domain-specific machine learning or artificial intelligence trustworthiness quality.

The cross-domain trustworthy machine learning or artificial intelligence configuration or delegation request may also include at least one of the following: an expected fairness parameter configured to indicate a relative fairness level for a domain-specific machine learning or artificial intelligence pipeline; an expected explainability parameter configured to indicate an expected explainability level for a domain-specific machine learning or artificial intelligence pipeline; an expected technical robustness parameter configured to indicate an expected technical robustness level for a domain-specific machine learning or artificial intelligence pipeline; and an expected adversarial robustness parameter configured to indicate an expected adversarial robustness level for a domain-specific machine learning or artificial intelligence pipeline.

Facilitating cross-domain network service-related machine learning or artificial intelligence pipeline trustworthiness functions may include, where the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network including at least two domains: generating a cross-domain trustworthiness machine learning or artificial intelligence capability information request and passing it to at least one domain-specific machine learning or artificial intelligence trustworthiness function, the cross-domain trustworthiness machine learning or artificial intelligence capability information request being configured to control at least one domain-specific machine learning or artificial intelligence trustworthiness function to enable capability discovery of the machine learning or artificial intelligence pipeline for at least one of the at least two domains; and obtaining a cross-domain trustworthiness machine learning or artificial intelligence capability information response from the at least one domain-specific machine learning or artificial intelligence trustworthiness function, reporting capability discovery for the machine learning or artificial intelligence pipeline for at least one of the at least two domains.

A cross-domain trustworthy machine learning or artificial intelligence capability information request may include: a domain scope parameter configured to identify the domain to which the request is being addressed; and a scope parameter configured to identify a domain-specific machine learning or artificial intelligence pipeline to which the request is being addressed.

The cross-domain trustworthiness machine learning or artificial intelligence capability information request may also include a pipeline stage parameter, which is configured to identify the stage of the domain-specific machine learning or artificial intelligence pipeline that the request is addressing.

Facilitating machine learning or artificial intelligence pipeline credibility functions associated with cross-domain network services may also include, where the cross-domain machine learning or artificial intelligence pipeline is used to control a cognitive autonomous network including at least two domains: obtaining a cross-domain credibility artificial intelligence report request or subscription from a network operator; based on the cross-domain credibility artificial intelligence report request or subscription from the network operator, generating a domain-specific credibility artificial intelligence report request or subscription and passing it to at least one domain-specific machine learning or artificial intelligence pipeline credibility function, wherein the domain-specific credibility artificial intelligence report request or subscription is configured to control at least one domain-specific machine learning or artificial intelligence pipeline credibility function to provide at least one domain-specific machine learning or artificial intelligence pipeline report response or notification; receiving machine learning or artificial intelligence capability information and/or at least one domain-specific machine learning or artificial intelligence pipeline report response or notification from at least one domain-specific machine learning or artificial intelligence pipeline credibility function; storing the machine learning or artificial intelligence capability information and/or at least one domain-specific machine learning or artificial intelligence pipeline report received from at least one domain-specific machine learning or artificial intelligence pipeline in a cross-domain trust knowledge database; generating and passing a cross-domain credibility artificial intelligence report based on at least one domain-specific machine learning or artificial intelligence pipeline report response or notification.

A cross-domain credibility AI report request may include: a domain scope parameter configured to identify the domain to which the request is being addressed; a scope parameter configured to identify a domain-specific machine learning or AI pipeline to which the request is being addressed; and a pipeline stage parameter configured to identify a stage of a domain-specific machine learning or AI pipeline to which the request is being addressed. The cross-domain trustworthy AI report request may also include at least one of the following: a list of fairness metric parameters, configured to identify fairness metrics to be reported; a list of fairness metric explanations, configured to identify fairness metric explanations to be reported; a list of explainability metrics, configured to identify explainability metrics to be reported; a list of explanations, configured to identify explanations to be reported; a list of technical robustness metrics, configured to identify technical robustness metrics to be reported; a list of technical robustness metric explanations, configured to identify technical robustness metric explanations to be reported; a list of adversarial robustness metrics, configured to identify adversarial robustness metrics to be reported; a list of adversarial robustness metric explanations, configured to identify adversarial robustness metric explanations to be reported; a start time parameter, configured to identify a start time for reporting; an end time parameter, configured to identify an end time for reporting; and a reporting interval parameter, configured to identify a periodic interval for reporting.

A cross-domain credibility AI report subscription may include: a domain scope parameter configured to identify the domain that the subscription is addressing; a scope parameter configured to identify the domain-specific machine learning or AI pipeline that the subscription is addressing; and a pipeline stage parameter configured to identify the stage of the domain-specific machine learning or AI pipeline that the subscription is addressing.

The cross-domain trustworthy AI report subscription may also include at least one of the following: a list of fairness metric parameters, configured to identify fairness metrics to be reported; a list of explainability metrics, configured to identify explainability metrics to be reported; a list of technical robustness metrics, configured to identify technical robustness metrics to be reported; a list of adversarial robustness metrics, configured to identify adversarial robustness metrics to be reported; and a cross-reporting threshold parameter, configured to identify a threshold against which a metric or metric interpretation is reported.

According to a third aspect, a device is provided, comprising at least one processor and at least one memory, the at least one memory comprising computer code for one or more programs, the at least one memory and the computer code together with the at least one processor being configured to enable the device to at least: facilitate machine learning or artificial intelligence pipeline credibility functions related to cross-domain network services, wherein the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network comprising at least two domains.

The device caused to facilitate cross-domain network service-related machine learning or artificial intelligence pipeline credibility functions can be caused to facilitate at least one of the following, wherein the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network comprising at least two domains: defining cross-domain network service-related machine learning or artificial intelligence pipeline credibility; configuring cross-domain network service-related machine learning or artificial intelligence pipeline credibility; measuring cross-domain network service-related machine learning or artificial intelligence pipeline credibility; and reporting machine learning or artificial intelligence pipeline credibility related to cross-domain network services.

The machine learning or artificial intelligence pipeline credibility function associated with cross-domain network services may include at least one of the following: fairness; explainability; and robustness.

The device caused to facilitate cross-domain network service-related machine learning or artificial intelligence pipeline credibility functions can be caused to: obtain cross-domain machine learning or artificial intelligence quality credibility, and the cross-domain machine learning or artificial intelligence quality credibility is configured to: cover domain-specific machine learning or artificial intelligence quality credibility requirements and constraints of cross-domain network service-related machine learning or artificial intelligence pipelines; convert the cross-domain machine learning or artificial intelligence quality credibility into at least one domain-specific machine learning or artificial intelligence quality credibility for at least one domain of the at least two domains; and provide at least one domain-specific machine learning or artificial intelligence quality credibility for at least one domain of the at least two domains.

The device, which is caused to convert the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality for at least one domain of at least two domains, can be caused to convert the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality based on the risk level of the cross-domain network service.

The device, which is caused to convert cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality for at least one domain of at least two domains, can be caused to convert the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality based on at least one service level agreement requirement for a cross-domain network, wherein at least one service level agreement includes at least one of the following: service type; service priority; and at least one key performance indicator metric.

The device, which is caused to provide at least one domain-specific machine learning or artificial intelligence credibility quality for at least one domain of at least two domains, can be caused to: generate a cross-domain credibility machine learning or artificial intelligence configuration or delegation request and pass it to at least one domain-specific machine learning or artificial intelligence credibility function, the cross-domain credibility machine learning or artificial intelligence configuration or delegation request being configured to control at least one domain-specific machine learning or artificial intelligence credibility function to configure a machine learning or artificial intelligence pipeline for at least one domain of at least two domains; and obtain a cross-domain credibility machine learning or artificial intelligence configuration or delegation response from at least one domain-specific machine learning or artificial intelligence credibility function based on the implementation of the configuration of the machine learning or artificial intelligence pipeline for at least one domain of at least two domains.

The device, which is caused to provide at least one domain-specific machine learning or artificial intelligence credibility quality for at least one domain of at least two domains, can be caused to: generate a cross-domain credibility machine learning or artificial intelligence configuration or delegation request and pass it to at least one domain-specific policy manager, the cross-domain credibility machine learning or artificial intelligence configuration or delegation request being configured to control at least one domain-specific machine learning or artificial intelligence credibility function to configure a machine learning or artificial intelligence pipeline for at least one domain of at least two domains; and obtain a cross-domain credibility machine learning or artificial intelligence configuration or delegation response from at least one domain policy manager based on the implementation of the configuration of the machine learning or artificial intelligence pipeline for at least one domain of at least two domains.

A cross-domain trustworthy machine learning or artificial intelligence configuration or delegation request may include: a domain scope parameter configured to identify the domain to which the request is being addressed; a pipeline identification parameter configured to identify the domain-specific machine learning or artificial intelligence pipeline to which the request is being addressed; and a category identification parameter configured to identify the at least one domain-specific machine learning or artificial intelligence trustworthiness quality.

The cross-domain trustworthy machine learning or artificial intelligence configuration or delegation request may also include at least one of the following: an expected fairness parameter configured to indicate a relative fairness level for a domain-specific machine learning or artificial intelligence pipeline; an expected explainability parameter configured to indicate an expected explainability level for a domain-specific machine learning or artificial intelligence pipeline; an expected technical robustness parameter configured to indicate an expected technical robustness level for a domain-specific machine learning or artificial intelligence pipeline; and an expected adversarial robustness parameter configured to indicate an expected adversarial robustness level for a domain-specific machine learning or artificial intelligence pipeline.

The device caused to facilitate cross-domain network service-related machine learning or artificial intelligence pipeline trustworthiness functions can be caused to: generate a cross-domain trustworthiness machine learning or artificial intelligence capability information request and pass it to at least one domain-specific machine learning or artificial intelligence trustworthiness function, the cross-domain trustworthiness machine learning or artificial intelligence capability information request being configured to control at least one domain-specific machine learning or artificial intelligence trustworthiness function to enable capability discovery of the machine learning or artificial intelligence pipeline for at least one of the at least two domains; and obtain a cross-domain trustworthiness machine learning or artificial intelligence capability information response from at least one domain-specific machine learning or artificial intelligence trustworthiness function, reporting capability discovery for the machine learning or artificial intelligence pipeline for at least one of the at least two domains.

A cross-domain trustworthy machine learning or artificial intelligence capability information request may include: a domain scope parameter configured to identify the domain to which the request is being addressed; and a scope parameter configured to identify a domain-specific machine learning or artificial intelligence pipeline to which the request is being addressed.

The cross-domain trustworthiness machine learning or artificial intelligence capability information request may also include a pipeline stage parameter, which is configured to identify the stage of the domain-specific machine learning or artificial intelligence pipeline that the request is addressing.

The device caused to facilitate machine learning or artificial intelligence pipeline credibility functions related to cross-domain network services may also be caused to: obtain a cross-domain credibility artificial intelligence report request or subscription from a network operator; based on the cross-domain credibility artificial intelligence report request or subscription from the network operator, generate a domain-specific credibility artificial intelligence report request or subscription and pass it to at least one domain-specific machine learning or artificial intelligence pipeline credibility function, wherein the domain-specific credibility artificial intelligence report request or subscription is configured to control at least one domain-specific machine learning or artificial intelligence pipeline credibility function to provide at least one domain-specific machine learning or artificial intelligence pipeline report response or notification; receive machine learning or artificial intelligence capability information and/or at least one domain-specific machine learning or artificial intelligence pipeline report response or notification from at least one domain-specific machine learning or artificial intelligence pipeline credibility function; store the machine learning or artificial intelligence capability information and/or at least one domain-specific machine learning or artificial intelligence pipeline report received from at least one domain-specific machine learning or artificial intelligence pipeline in a cross-domain trust knowledge database; generate and pass a cross-domain credibility artificial intelligence report based on at least one domain-specific machine learning or artificial intelligence pipeline report response or notification.

A cross-domain credibility AI report request may include: a domain scope parameter configured to identify the domain to which the request is being addressed; a scope parameter configured to identify a domain-specific machine learning or AI pipeline to which the request is being addressed; and a pipeline stage parameter configured to identify a stage of a domain-specific machine learning or AI pipeline to which the request is being addressed.

The cross-domain trustworthy AI report request may also include at least one of the following: a list of fairness metric parameters, configured to identify fairness metrics to be reported; a list of fairness metric explanations, configured to identify fairness metric explanations to be reported; a list of explainability metrics, configured to identify explainability metrics to be reported; a list of explanations, configured to identify explanations to be reported; a list of technical robustness metrics, configured to identify technical robustness metrics to be reported; a list of technical robustness metric explanations, configured to identify technical robustness metric explanations to be reported; a list of adversarial robustness metrics, configured to identify adversarial robustness metrics to be reported; a list of adversarial robustness metric explanations, configured to identify adversarial robustness metric explanations to be reported; a start time parameter, configured to identify a start time for reporting; an end time parameter, configured to identify an end time for reporting; and a reporting interval parameter, configured to identify a periodic interval for reporting.

A cross-domain credibility AI report subscription may include: a domain scope parameter configured to identify the domain that the subscription is addressing; a scope parameter configured to identify the domain-specific machine learning or AI pipeline that the subscription is addressing; and a pipeline stage parameter configured to identify the stage of the domain-specific machine learning or AI pipeline that the subscription is addressing.

The cross-domain trustworthy AI report subscription may also include at least one of the following: a list of fairness metric parameters, configured to identify fairness metrics to be reported; a list of explainability metrics, configured to identify explainability metrics to be reported; a list of technical robustness metrics, configured to identify technical robustness metrics to be reported; a list of adversarial robustness metrics, configured to identify adversarial robustness metrics to be reported; and a cross-reporting threshold parameter, configured to identify a threshold against which a metric or metric interpretation is reported.

According to a fourth aspect, an apparatus is provided, comprising a circuit system configured to facilitate machine learning or artificial intelligence pipeline credibility functions related to cross-domain network services, wherein the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network including at least two domains.

According to a fifth aspect, a computer program is provided, comprising computer executable code, which, when executed on at least one processor, is configured to: facilitate machine learning or artificial intelligence pipeline credibility functions related to cross-domain network services, wherein the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network comprising at least two domains.

According to a sixth aspect, there is provided a computer program comprising instructions [or a computer-readable medium comprising program instructions] for causing an apparatus to perform at least the following: facilitating machine learning or artificial intelligence pipeline credibility functions associated with cross-domain network services, wherein the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network comprising at least two domains.

According to a seventh aspect, a non-transitory computer-readable medium is provided, comprising program instructions for causing a device to perform at least the following: facilitating machine learning or artificial intelligence pipeline credibility functions related to cross-domain network services, wherein the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network comprising at least two domains.

According to an eighth aspect, a non-transitory computer-readable medium is provided, comprising program instructions for causing a device to perform at least the following: facilitating machine learning or artificial intelligence pipeline credibility functions related to cross-domain network services, wherein the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network comprising at least two domains.

According to a ninth aspect, a computer-readable medium is provided, comprising program instructions for causing an apparatus to perform at least the following: facilitating machine learning or artificial intelligence pipeline credibility functions associated with cross-domain network services, wherein the cross-domain machine learning or artificial intelligence pipeline is used for control of a cognitive autonomous network comprising at least two domains.

An apparatus comprises means for performing the actions of the method as described above.

A device is configured to perform the actions of the method described above.

A computer program includes program instructions for causing a computer to execute the method described above.

A computer program product stored on a medium can cause an apparatus to execute the method described herein.

An electronic device may include the apparatus described herein.

A chipset may include the device described herein.

The embodiments of the present application are intended to solve the problems associated with the prior art.

In the above, many different aspects have been described. It should be understood that additional aspects can be provided by combining any two or more of the above aspects.

Various other aspects are also described in the following detailed description and appended claims.

Abbreviations

AI: Artificial Intelligence

CAN: Cognitive Autonomous Network

CD: Cross-domain

CN: Core Network

CNF: Cognitive Network Function

CRUD: Create, Read, Update, Delete

CDSMD: Cross-domain Service Management Domain

CU: Centralized Unit

DU: Distributed Unit

E2E: End to End

E2ESMD: End-to-End Service Management Domain

HLEG: High Level Expert Group

MANO: Management and Orchestration

MD: Management Domain

ML: Machine Learning

QCI: QoS Class Identifier

QoE: Quality of Experience

QoS: Quality of Service

QoT: Quality of Trustworthiness

RAN: Radio Access Network

RRU: Remote Radio Unit

SLA: Service Level Agreement

TAI: Trusted Artificial Intelligence

TAIF: TAI Framework

TN：Transmission Network

UMTS: Universal Mobile Telecommunications System

URLLC: Ultra-Reliable Low Latency Communications

VNF: Virtual Network Function

V2X: Vehicle to Everything

WI: Work Item

3GPP: Third Generation Partnership Project

5G: Fifth Generation

5GC: 5G Core Network

5GS: 5G System

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments will now be described, by way of example only, with reference to the accompanying drawings, in which:

FIG1 shows a schematic diagram of a 5G communication system;

FIG2 shows a schematic diagram of a control device;

FIG3 shows a schematic diagram of a terminal;

FIG4 shows a schematic diagram of an artificial intelligence/machine learning pipeline;

FIG5 shows a schematic diagram of an example of a trusted artificial intelligence framework for a cognitive autonomous network;

FIG6 shows a workflow diagram for an example trusted artificial intelligence framework for a cognitive autonomous network as shown in FIG5 ;

FIG7 illustrates a schematic diagram of an example cross-domain management and orchestration architecture utilizing a domain-specific trusted artificial intelligence framework;

FIG8 illustrates a schematic diagram of an example cross-domain management and orchestration architecture utilizing a domain-specific trusted artificial intelligence framework according to some embodiments;

FIG9 illustrates a workflow diagram of an example cross-domain trusted artificial intelligence application programming interface provided by a domain-specific artificial intelligence trust engine to a cross-domain artificial intelligence trust engine; and

FIG. 10 shows a schematic diagram of a non-volatile storage medium storing instructions that, when executed by a processor, allow the processor to perform one or more steps of the methods described herein.

DETAILED DESCRIPTION

Certain embodiments are explained below with reference to a mobile communication device capable of communicating via a wireless cellular system and a mobile communication system serving such a mobile communication device. Before explaining the exemplary embodiments in detail, certain general principles of a wireless communication system, its access system, and a mobile communication device are briefly explained with reference to Figures 1, 2, and 3 to help understand the technology behind the described examples.

Figure 1 shows a schematic diagram of a 5G system (5GS). The 5GS may include a terminal, a (radio) access network ((R)AN), a 5G core network (5GC), one or more application functions (AFs), and one or more data networks (DNs).

The 5G(R)AN may include one or more gNodeB (gNB) distributed unit (DU) functions connected to one or more gNodeB (gNB) centralized unit (CU) functions.

5GC may include access and mobility management function (AMF), session management function (SMF), authentication server function (AUSF), user data management (UDM), user plane function (UPF), network exposure function (NEF) and/or other network functions (NFs) not represented, such as operation administration and maintenance (OAM) NF.

FIG2 illustrates an example of a control device 200 for controlling the functions of a (R)AN or 5GC as shown in FIG1 . The control device may include at least one random access memory (RAM) 211a, at least one read-only memory (ROM) 211b, at least one processor 212, 213, and an input/output interface 214. At least one processor 212, 213 may be coupled to the RAM 211a and the ROM 211b. At least one processor 212, 213 may be configured to execute appropriate software code 215. The software code 215 may, for example, allow one or more steps to be executed to perform one or more of the present aspects. The software code 215 may be stored in the ROM 211b. The control device 200 may be interconnected with another control device 200 that controls another function of a 5G (R)AN or 5GC. In some embodiments, each function of the (R)AN or 5GC includes a control device 200. In an alternative embodiment, two or more functions of the (R)AN or 5GC may share a control device.

FIG3 shows an example of a terminal 300, such as the terminal shown in FIG1. The terminal 300 may be provided by any device capable of sending and receiving radio signals. Non-limiting examples include user equipment, a mobile station (MS) or a mobile device (such as a mobile phone or so-called "smart phone"), a computer provided with a wireless interface card or other wireless interface facilities (e.g., a USB dongle), a personal data assistant (PDA) or a tablet computer provided with wireless communication capabilities, a machine type communication (MTC) device, a cellular Internet of Things (CIoT) device, or any combination of these devices, etc. The terminal 300 may provide, for example, communication for carrying data for communication. The communication may be one or more of voice, email (email), text message, multimedia, data, machine data, etc.

The terminal 300 may receive signals via an air or radio interface 307 via appropriate means for receiving, and may send signals via appropriate means for sending radio signals. In FIG. 3 , the transceiver means is schematically designated by a block 306. The transceiver means 306 may be provided, for example, by a radio part and an associated antenna arrangement. The antenna arrangement may be arranged internally or externally to the mobile device.

The terminal 300 may be provided with at least one processor 301, at least one memory ROM 302a, at least one RAM 302b and other possible components 303 for use in software and hardware assisted execution of the tasks it is designed to perform, including control of access to and communication with access systems and other communication devices. At least one processor 301 is coupled to RAM 302b and ROM 302a. At least one processor 301 may be configured to execute appropriate software code 308. Software code 308 may, for example, allow execution of one or more of the current aspects. Software code 308 may be stored in ROM 302a.

The processor, memory and other related control means may be provided on an appropriate circuit board and/or in a chipset. This feature is indicated by reference numeral 304. The device may optionally have a user interface, such as a keypad 305, a touch sensitive screen or pad, a combination thereof, etc. Optionally, one or more of a display, a speaker and a microphone may be provided depending on the type of device.

As mentioned earlier, closed-loop automation and machine learning (ML) (also known as artificial intelligence or AI) can be built into self-organizing networks (SON) or cognitive autonomous networks (CAN), enabling operators to automatically optimize each cell in the radio access network.

Artificial Intelligence (AI) or Machine Learning (ML) pipelines help automate AI/ML workflows by breaking them down into independent, reusable, and modular components that are then connected together through pipelines to create models. AI/ML pipelines are iterative, where each step is repeated to continually improve the accuracy of the model.

As shown in Figure 4, the example AI/ML workflow includes the following three components:

Data source manager 403. The data source manager 403 is configured to implement functions such as data collection and data preparation.

Model training manager 405. Model training manager 405 is configured to implement functions such as hyperparameter tuning.

Model reasoning manager 407. Model reasoning manager 407 is configured to implement functions such as model evaluation.

With AI/ML pipelineization and the recent push for microservice architectures (e.g., containers), each AI/ML workflow component is abstracted as an independent service that can be handled independently by relevant stakeholders (e.g., data engineers, data scientists). In addition, an AI/ML pipeline orchestrator 401 (an example of which is provided by Kubeflow) can manage the lifecycle of the AI/ML pipeline. For example, it manages the various stages in the debugging, scaling, and decommissioning lifecycles.

In order for AI/ML systems to gain widespread acceptance, in addition to their performance (e.g. accuracy), they should also be trustworthy. Legal bodies are proposing frameworks on AI/ML applications, for example the European Commission has proposed the first ever legal framework on AI. The legal framework lays out new rules for AI to be trustworthy and for mission-critical AI-based systems to comply with in the near future. The High-Level Expert Group (HLEG) group on AI has developed the European Commission’s Trustworthy AI (TAI) strategy. In the deliverable “Ethics Guidelines for Trustworthy AI” published in April 2019, the group has listed seven key requirements that an AI system should meet to be considered trustworthy. Here are the requirements:

Transparency. Transparency requirements include traceability, explainability, and communication.

Diversity, non-discrimination, and equity. This requirement includes avoiding unfair bias, accessibility and universal design, and stakeholder engagement.

Technical robustness and security. This requirement includes resistance to attacks and security, fallback plans, and general safety, accuracy, reliability, and reproducibility.

Privacy and data governance. This requirement includes respect for privacy, data quality and integrity, and access to data.

Accountability. Accountability requirements include auditability, minimization and reporting of adverse impacts, trade-offs, and remediation.

Human agency and oversight. Human agency and oversight requirements include fundamental rights, human agency, and human oversight.

Social and environmental well-being. This requirement includes sustainability and environmental friendliness, social impact, society and democracy.

In addition, the International Organization for Standardization and the International Electrotechnical Commission (ISO/IEC) also released a technical report on "Overview of trustworthiness in artificial intelligence". Early efforts in the open source community are also reflected in the development of TAI frameworks/tools/libraries, such as IBM AI360, Google Expandable AI, and TensorFlow Responsible AI.

Below, we present some key TAI definitions/algorithms/metrics described in the AI/ML research community.

Fairness: Fairness is the process of understanding the bias introduced in the data and ensuring that the model provides fair predictions across all demographic groups. It is important to apply fairness analysis throughout the AI/ML pipeline, ensuring that models are constantly re-evaluated from a fairness and inclusion perspective. This is especially important when AI/ML is deployed in critical business processes that impact a wide range of end users. There are three broad approaches to detecting bias in AI/ML models:

1. Pre-processing fairness - using algorithms (such as reweighting and disparate influence removers) to detect bias in AI/ML training data.

2. Processing fairness - using algorithms (such as bias removers and adversarial debiasing) to detect bias in AI/ML model generation.

3. Post-processing fairness – using algorithms (such as odds equalization and rejection option classification) to detect bias in AI/ML model decision making.

Quantification of fairness - Several metrics exist to measure individual and group fairness. For example, statistical parity difference, mean odds difference, differential impact, and the Theil index.

Explainability: Explainability of AI/ML models refers to the demystification of a black box model that only makes predictions or gives recommendations to a white box that actually gives details of the underlying mechanisms and patterns identified by the model for a specific dataset. There are multiple reasons why it is necessary to understand the underlying mechanisms of AI/ML models such as human readability, reasonableness, interpretability, and bias mitigation. There are three broad approaches to designing explainable ML models:

1. Pre-modeling interpretability - understanding or describing the data used to develop AI/ML models. For example, using algorithms such as ProtoDash and Disentangled Inference Prior Variational Autoencoder Explainer.

2. Explainable Modeling/Explanable Modeling - Develop more explainable AI/ML models, such as ML models with joint prediction and explanation or alternative explainable models. For example, using algorithms such as generalized linear rule models and Teaching Explainable Decisions (TED).

3. Post-modeling explainability - extract explanations from pre-developed AI/ML models. For example, using algorithms such as ProtoDash, Contrastive Explanations Method, Profweight, LIME, and SHAP.

Furthermore, explanations can be local (i.e. explaining a single instance/prediction) or global (i.e. explaining the global AI/ML model structure/prediction, e.g. based on combining many local explanations for each prediction).

Quantification of Interpretability — While it is ultimately up to the consumer to determine the quality of an explanation, the research community has proposed quantitative metrics as proxies for interpretability. Several metrics exist to measure interpretability, such as faithfulness and monotonicity.

Robustness (Adversarial): There are four adversarial threats that any AI/ML model developer/scientist needs to consider defending and evaluating their AI/ML models and applications.

1. Evasion: Evasion attacks involve carefully perturbing input samples at test time so that they are misclassified. For example, using techniques such as shadow attacks and threshold attacks.

2. Poisoning: Poisoning is the adversarial contamination of training data. Machine learning systems can be retrained using data collected during operation. An attacker may poison this data by injecting malicious samples during operation, thereby undermining the retraining. For example, using techniques such as backdoor attacks and adversarial backdoor embedding.

3. Extraction: Extraction attacks aim to replicate machine learning models by querying the target model, for example, using techniques such as KnockoffNets and functional equivalent extraction.

4. Inference: Inference attacks determine whether a sample of data was used in the training dataset of an AI/ML model. For example, using techniques such as membership inference black box and attribute inference black box.

In addition to adversarial robustness, there are other aspects of AI/ML technology robustness that need to be addressed, such as handling missing data, erroneous data, data confidence, backup plans, etc.

There are multiple ways to protect AI/ML models from such adversarial attacks at every stage of AI/ML design:

Preprocessors — For example, using techniques such as InverseGAN and DefenseGAN.

Post-processing — for example, using techniques such as inverse sigmoid and rounding.

Trainer — For example, using techniques such as general adversarial training and Madry’s protocol.

Transformers – For example, using techniques such as Defense Distillation and Neural Purification.

Detectors - for example, using techniques such as activation analysis based detection and spectral signature based detection.

Quantification of robustness: There are several metrics to measure the robustness of ML models, such as empirical robustness and loss sensitivity.

An example of a Trusted Artificial Intelligence Framework (TAIF) for Cognitive Autonomous Networks (CAN) to facilitate the definition, configuration, monitoring and measurement of AI/ML model trustworthiness (e.g., fairness, explainability, robustness) in interoperable and multi-supplier environments is shown in FIG5 , and further details can be found in PCT/EP2021/062396 .

In this example, a network operator 501 is shown, which is configured to pass information to a policy manager 533 and an AI trust engine 503.

The policy manager 533 is configured to receive information from the network operator 501. In addition, the policy manager 533 is configured to receive or otherwise obtain a service definition or business/customer intent. In addition to network/AI quality of service (QoS) requirements, the service definition or business/customer intent may also include AI/ML trustworthiness requirements, and TAIF is used to configure the requested AI/ML trustworthiness and monitor and ensure its implementation.

Thus, for example, the system may include a service management and orchestration 527 function configured to receive quality of service (QoS) from a policy manager 533. The service management and orchestration 527 is configured to control an element manager or a virtual network function (VNF) manager or a resource manager 531 based on the output of the service management and orchestration function 527.

In addition, the system may include an AI pipeline orchestrator 525. The AI pipeline orchestrator 525 is configured to obtain or receive AI QoS from the policy manager 533, and based thereon, and in a similar manner as described with respect to FIG. 4, is configured to control the operation of the data source managers 509, 519, the model training managers 511, 521, and the model inference managers 513, 523 for the AI pipeline 1 505 and the AI pipeline 2 515.

TAIF introduces two additional management functions, the AI Trust Engine (Trustworthiness Function) 503 (one per management domain) and the AI Trust Manager 507, 617 (one per AI/ML pipeline 505, 515) and six new interfaces (T1-T6) configured to support interactions in TAIF. The AI Trust Engine 503 is configured to serve as a hub for managing all AI trustworthiness related components in the network, while the AI Trust Managers 507, 517 are use case and typically vendor specific, with knowledge of the AI use case and how it is implemented.

In addition, the example TAIF also adopts the concept of AI Trust Quality (AIQoT) to define AI/ML model trustworthiness in a unified way, covering three factors, namely fairness, explainability and robustness. AIQoT is passed from the policy manager 533 to the AI trust engine function 503, for example, and is similar to how QoS is used for network performance.

Example QoT can be shown by the following table

A high-level general workflow in the example TAIF is shown in Figure 6.

In this example workflow, customer intent is shown being provided to the policy manager function 533, as shown in step 601 in FIG. 6 .

Also shown is the network operator (via the policy manager function 533 ) specifying the desired AIQoT (use case specific based on risk level) to the AI trust engine 503 , for example, via a T1 interface, as shown by step 603 in FIG. 6 .

The AI trust engine 503 converts the AIQoT into specific AI trust (i.e., fairness, explainability, and robustness) requirements and identifies the affected use case(s) specific AI trust manager. Using the T2 interface, the AI trust engine 503 configures the AI trust manager 507, as shown in FIG6 by step 605.

The use case specific and implementation-aware AI trust manager 507 is configured to configure, monitor and measure AI trust requirements for the AI data source manager 509 through the T3 interface, as shown by step 607 in Figure 6.

Furthermore, the use case specific and implementation-aware AI trust manager 507 is configured to configure, monitor and measure AI trust requirements for the AI training manager 511 via the T4 interface, as shown by step 609 in FIG. 6 .

Additionally, the use case specific and implementation-aware AI trust manager 507 is configured to configure, monitor and measure AI trust requirements for the AI reasoning manager 513 via the T5 interface, as shown in FIG. 6 by step 611 .

The measured or collected TAI metrics and/or TAI interpretations about the AI pipeline from the AI data source manager 509, the AI training manager 511, and the AI reasoning manager 513 are pushed to the AI trust manager 507 through the T3, T4, and T5 interfaces, respectively, as shown in FIG. 6 by steps 613, 615, and 617, respectively.

The AI trust manager 507 pushes the TAI measurements and/or TAI interpretations to the AI trust engine 503 through the T2 interface based on the reporting mechanism configured by the AI trust engine (trustworthiness function), as shown by step 619 in Figure 6.

Finally, the network operator 501 may request (as shown by step 621 in FIG. 6 ) and receive (as shown by step 623 in FIG. 6 ) TAI metrics/explanations of the AI pipeline from the AI trust engine via the T6 interface.

Based on the acquired information, the network operator may decide to update the policy via the policy/intent manager, as shown in FIG. 6 by step 625 .

Thus, the example TAI framework enables various telecom stakeholders (e.g., cognitive network function providers, network operators, regulators, end users) to trust the decisions/predictions made by AI/ML models in the network.

An example cross-domain management and orchestration architecture is also shown in Figure 7. This example shows a cross-domain end-to-end (E2E) network service scenario, but other cross-domain non-E2E scenarios (i.e., within each domain) are possible. For example, the core domain can recursively embed the 3GPP-defined network function (NF) domain and virtualization domain, and the RAN domain can include centralized units (CU), distributed units (DU), remote radio units (RRU), midhaul and fronthaul domains provided by different suppliers.

In the example cross-domain E2E network service scenario shown in FIG7 , a cross-domain service management domain (CDSMD) (e.g., E2E service management domain) 705 is located between a cross-domain policy/intent manager 703 and a network operator 701, and is configured to control the cross-domain service MD 705 and the following domain management domains. For example, as shown in FIG7 , a (first) domain 1 MD (e.g., RAN MD) 712, a (second) domain 2 MD (e.g., transport MD) 722, and a (third) domain 3 MD (e.g., core network MD) 732 are shown. The cross-domain service management domain (CDSMD) 705 is configured to decompose a cross-domain E2E network service request (according to a service level agreement (SLA)) received from a network operator 701 or a customer (e.g., via a cross-domain policy/intent manager 703) into domain-specific (e.g., RAN, transport, core) network resource/service requirements, and transmit them to the corresponding separate management domains (MDs) 712, 722, 732.

The individual MDs 712 , 722 , 732 are also configured to be responsible for ensuring that domain-specific resource/service requirements are met within their corresponding domains by continuously monitoring and reporting resource/service-related key performance indicators (KPIs) to the CDSMD 705 .

In the example shown in FIG. 7 , a (first) domain 1 MD (e.g., RAN MD) 712 is shown. The (first) domain 1 MD (e.g., RAN MD) 712 may also be controlled/monitored by a domain-specific policy/intent manager 711 (which is similar to the policy/intent manager shown in FIGS. 5 and 6 ). The (first) domain 1 MD (e.g., RAN MD) 712 may include a domain 1 AI trust engine 713 (or trustworthiness function) and a domain 1 AI pipeline orchestrator 714, both of which manage and control the domain 1 AI pipeline 1 715. The 1 AI pipeline 1 715 may then control or configure domain-specific aspects of cross-domain network services 717, domain 1 resources (RAN resources 716).

In addition, the example shown in FIG7 shows a (second) domain 2 MD (e.g., transport MD) 722. The (second) domain 2 MD 722 may also be controlled/monitored by a domain-specific policy/intent manager 721 (which is similar to the policy/intent manager shown in FIGS. 5 and 6). The (second) domain 2 MD 722 may include a domain 2 AI trust engine 723 (or trustworthiness function) and a domain 2 AI pipeline orchestrator 724, both of which manage and control the domain 2 AI pipeline 2 725. The domain 2 AI pipeline 2 725 may then control or configure domain-specific aspects of the cross-domain network service 717, domain 2 resources (transport resources 726).

In addition, as shown in the example of FIG. 7 , a (third) domain 3MD (e.g., core network MD) 732 may be employed. The (third) domain 3MD 732 may also be controlled/monitored by a domain-specific policy/intent manager 731 (similar to the policy/intent manager shown in FIGS. 5 and 6 ). The (third) domain 3MD 732 may include a domain 3 AI trust engine 733 (or trustworthiness function) and a domain 3 AI pipeline orchestrator 734, both of which manage and control the domain 3 AI pipeline 3 735. The domain 3 AI pipeline 3 735 may then control or configure domain-specific aspects of the cross-domain network service 717, domain 3 resources (core network resources 736).

Thus, the requested/instantiated cross-domain E2E network service (e.g., covering RAN, transport and core domains (represented by box 717)) is managed by its corresponding AI pipeline (or CNF) in the corresponding MD (represented by boxes 715, 725, 735). It should be noted that, depending on the use case, the AI pipeline can be instantiated in a domain-specific MD (e.g., for proactive resource autoscaling) or within the domain itself (e.g., for proactive mobility switching in the RAN domain). Then, utilizing the domain-specific AI trust engine (trustworthiness function) and the AI pipeline's specific AI trust manager (as described and referenced above), the AI pipeline trustworthiness (i.e., AIQoT including fairness, explainability, robustness) for the domain-specific AI pipeline can be defined, configured, measured and reported within the corresponding MD.

However, the network shown in FIG7 may have problems because there is no management function configured to receive the desired cross-domain AIQoT (i.e., defined by the cross-domain policy/intent manager based on the risk level of the cross-domain E2E network service) for the cross-domain E2E network service in CDSMD 703. Therefore, CDSMD has no way to:

Convert cross-domain AIQoT into domain-specific AIQoT;

Discover TAI capability information from domain-specific (multiple) AI trust engines;

delivering the converted domain-specific AIQoT to domain-specific AI trust engine(s); and

Collect/request cross-domain TAI metrics/explanations from (multiple) domain-specific AI trust engines.

Furthermore, even if this issue is addressed, CDSMD is unable to resolve (e.g., perform root cause analysis) TAI-related escalations (i.e., in terms of TAI measurements/explanations) received from (multiple) domain-specific AI trust engines that pertain to cross-domain E2E network services, and is unable to delegate information related to relevant TAI escalations received from the domain-specific AI trust engine of one MD to another MD so that the other MD can take preventive measures to avoid cross-domain E2E network service SLA violations (in our case, cross-domain AIQoT).

Additionally, such a system does not have a way to aggregate TAI-related escalation metrics/explanations received from the (multiple) AI trust engines of separate MDs to provide a CDSMD to provide a global view of the problem to the network operator or customer. In this example, TAI-related escalation metrics/explanations may include cross-domain AIQoT violations.

A concept further discussed in the following embodiments is the introduction of a cross-domain TAI framework (which extends the domain-specific TAI framework discussed above and introduced in PCT/EP2021/062396 for cognitive autonomous networks to facilitate the definition, configuration, measurement and reporting of AI pipeline trustworthiness (i.e. fairness, explainability, robustness) related to cross-domain network services in interoperable and multi-supplier environments.

In these embodiments, in addition to the cross-domain QoS requirements, the customer intent corresponding to the network service may also include cross-domain AI trust requirements, and the cross-domain TAI framework is used to ensure that the desired cross-domain AI trust requirements are met.

As shown in Figure 8, the cross-domain TAI framework includes a novel management function, a cross-domain AI trust engine (credibility function) 801 (or credibility function). In some embodiments, the cross-domain AI trust engine is adopted within the cross-domain service management domain (CDSMD). In addition, in some embodiments, a new interface (which is designated as TCD-1 in the example shown in Figure 8) 805 is implemented, which is configured to support the interaction between the cross-domain AI trust engine 801 and the domain-specific AI trust engines 813, 823, 833 (or credibility functions). In addition, in some embodiments, another new interface (which is designated as TCD-2 in the example shown in Figure 8) is implemented between the cross-domain AI trust engine 801 and the domain-specific policy/intent manager 811, 821, 831.

Furthermore, in the embodiments described in detail herein, the concept of cross-domain AIQoT is introduced to define cross-domain AI trustworthiness in a unified manner, covering domain-specific AIQoT requirements and constraints for AI pipelines related to cross-domain network services.

In some embodiments, the cross-domain AI trust engine 801 is configured to support the following operations:

Depending on the risk level of cross-domain network services, convert cross-domain AIQoT requirements into domain-specific AIQoT requirements (e.g., RAN domain AIQoT, transport domain AIQoT, and core domain AI QoT);

The domain-specific AI trust engine enables discovery/determination of TAI methods and/or TAI metrics and/or TAI interpretations configured in a domain-specific AI pipeline belonging to a cross-domain network service;

Verification that cross-domain AIQoT and/or domain-specific AIQoT requirements for cross-domain network services are met;

Configuration/delegation of the desired/updated AIQoT (derived from the cross-domain AIQoT) that the domain-specific AI trust engine needs to satisfy in the domain-specific AI pipeline belonging to the cross-domain network service;

Requests/subscriptions for TAI metrics and/or TAI interpretations that the domain-specific AI trust engines 813, 823, 833 are able to measure/report/upgrade in domain-specific AI pipelines belonging to cross-domain network services;

storing all TAI capability information and/or TAI reports received from the domain-specific AI trust engines 813, 823, 833 regarding the domain-specific AI pipelines 715, 725, 735 belonging to the cross-domain network service in a cross-domain trust knowledge database;

Perform root cause analysis of TAI reports received from domain-specific AI trust engines 813, 823, 833. Also, update domain-specific AIQoT requirements based on TAI reports if needed;

Providing a global view to the network operator on issues/escalations of cross-domain network services (eg, aggregated cross-domain network service related TAI reports) - which in this example may be cross-domain AIQoT violations.

In some embodiments, a cross-domain TAI API may be implemented (which may be generated by (multiple) domain-specific AI trust engines and consumed by the cross-domain AI trust engine). For example, these may be the following:

1. Cross-domain TAI Capability Discovery API (Req/Resp) - The cross-domain TAI capability discovery API is configured to allow the cross-domain AI trust engine to discover, via the TCD-1 interface, TAI methods and/or TAI metrics and/or TAI interpretations that the domain-specific AI trust engine is capable of configuring in a domain-specific AI pipeline belonging to a cross-domain network service.

2. Cross-domain TAI configuration API or cross-domain TAI delegation API (Req/Resp). The cross-domain TAI configuration API or cross-domain TAI delegation API is configured to allow the cross-domain AI trust engine to configure/delegate the expected/updated AIQoT (derived from the cross-domain AIQoT) that needs to be met in the domain-specific AI pipeline belonging to the cross-domain network service to the domain-specific AI trust engine via the TCD-1 interface. Alternatively, in some embodiments, the cross-domain TAI configuration API or cross-domain TAI delegation API is configured to allow the cross-domain AI trust engine to notify the domain-specific policy/intent manager (via the domain-specific AI trust engine) via the TCD-2 interface of the expected/updated AIQoT (derived from the cross-domain AIQoT) that is required to be configured in the domain-specific AI pipeline belonging to the cross-domain network service.

3. Cross-domain TAI reporting API or cross-domain TAI upgrade API (request/response and subscription/notification). The cross-domain TAI reporting API or cross-domain TAI upgrade API is configured to allow the cross-domain AI trust engine to request/subscribe to the domain-specific AI trust engine via the TCD-1 interface to measure/report/upgrade TAI metrics and/or TAI interpretations in the domain-specific AI pipeline belonging to the cross-domain network service.

Figure 9 shows an example workflow diagram illustrating an implementation of a cross-domain AI trust engine 801 according to some embodiments. In addition, the figure also shows the application of the above-mentioned API and is supplied by (multiple) domain-specific AI trust engines through the TCD-1 interface to discover, configure, measure and query/collect TAI methods and/or TAI metrics and/or TAI interpretations from domain-specific AI pipelines belonging to cross-domain network services.

Additionally, an additional example alternative implementation is shown, showing the interaction between the cross-domain AI trust engine and the domain-specific policy/intent manager via the TCD-2 interface.

As shown by step 901 in FIG. 9 , the network operator 701 notifies the cross-domain policy/intent manager 703 (within the cross-domain service management domain 705 ) of the intent for cross-domain network services, for example, via a T1 interface.

Then, the cross-domain policy/intent manager 703 is configured to convert the intent for the cross-domain network service into a cross-domain AIQoT and notify the cross-domain AI trust engine 801 through step 903, as shown in Figure 9.

Then, the cross-domain AI trust engine 801 can be configured to convert the obtained cross-domain AIQoT requirements into domain-specific AIQoT requirements, as shown by step 905 in Figure 9. In other words, the cross-domain AI trust engine 801 is configured to generate parameters RAN domain AIQoT, transport domain AIQoT, and core domain AIQoT depending on the risk level of the cross-domain network service. The generation of the parameters RAN domain AIQoT, transport domain AIQoT, and core domain AIQoT depending on the risk level of the cross-domain network service can be shown, for example, by the following table:

In some embodiments, the conversion/mapping logic may consider SLA requirements (eg, service type, service priority, KPI metrics) for cross-domain network services, and optionally, domain-specific TAI capability information. In some embodiments, the conversion may be performed after step 911 .

The operations of steps 907 , 909 and 911 shown in FIG. 9 may be implemented as a part of a cross-domain TAI capability discovery API as shown by reference numeral 906 in FIG. 9 .

For example, the cross-domain AI trust engine 801 can be configured to generate a cross-domain TAI capability information request (CDTAICIREq). As shown in Figure 9, the request can be sent from the cross-domain AI trust engine 801 to the (multiple) domain-specific AI trust engine through step 907. In this example, the domain-specific AI trust engine is the RAN management domain 712 AI trust engine 813. The request is used to request information about the TAI method and/or TAI metric and/or TAI interpretation that the (multiple) AI trust engine 813 can configure in the domain-specific AI pipeline belonging to the cross-domain network service.

In some embodiments, an example request (CDTAICIReq) includes the following parameters:

In some embodiments, (multiple) domain-specific AI trust engines (in this example, RAN management domain 712 AI trust engine 813) are configured to determine all information requested in the received request (CDTAICIReq) by interacting with a domain-specific AI pipeline as shown by step 907 in Figure 9. The interaction in this example is with the RAN management domain 712, AI pipeline 715 (e.g., AI trust manager 902, AI data source manager 903, AI training manager 904, AI reasoning manager 905) belonging to the cross-domain network service. The details of the operation can be implemented in any suitable manner, for example as described in PCT/EP2021/071044.

In some embodiments, the domain-specific AI trust engine (in this example, the RAN management domain 712 AI trust engine 813) sends a cross-domain TAI capability information response (CDTAICIResp) back to the requesting cross-domain AI trust engine 801 as shown by step 911 in Figure 9. In some embodiments, the cross-domain TAI capability information response (CDTAICIResp) includes information about supported TAI methods and/or TAI metrics and/or TAI interpreted domain-specific AI pipelines (belonging to cross-domain network services).

In some embodiments, CDTAICIResp includes the following parameters:

In some embodiments, the cross-domain AI trust engine is configured to store the cross-domain TAI capability information in a cross-domain trust knowledge database. If there is a capability update in one of the domains, the database can be updated.

In some embodiments, the cross-domain AI trust engine 801 is configured to determine that cross-domain AIQoT and/or domain-specific AIQoT requirements can be met based on the cross-domain TAI capability information response as shown by step 913 in Figure 9 or the cross-domain TAI capability information stored in the cross-domain trust knowledge database. In addition, in some embodiments, the cross-domain AI trust engine 801 is configured to convert the domain-specific AIQoT into fairness, explainability and robustness requirements, as shown and described above, to determine the cross-domain AIQoT and/or domain-specific AIQoT satisfaction results.

The cross-domain AI trust engine 801 may also be configured to generate and send a negative acknowledgement (cross-domain AIQoT NACK) to the cross-domain policy/intent manager 703 as shown in step 915 in FIG. 9 if the requirement cannot be met. This is sent in response to the cross-domain AIQoT received from step 903.

The operations shown as steps 917 , 919 and 921 in FIG. 9 show a first alternative as part of a cross-domain TAI configuration API or a cross-domain TAI delegation API as shown by reference numeral 916 in FIG. 9 .

For example, a cross-domain TAI configuration/delegation (CRUD) request (CDTAIConReq) is sent from the cross-domain AI trust engine 801 to the domain-specific AI trust engine(s), which in this example is the RAN management domain 712 AI trust engine 813, to notify the converted domain-specific AIQoT that needs to be satisfied in the domain-specific AI pipeline belonging to the cross-domain network service. This is shown as step 917 in FIG. 9.

In some embodiments, the cross-domain AI trust engine 801 is configured to also convert domain-specific AIQoT into fairness, explainability, and robustness requirements, as shown and described above, and include this information in the cross-domain TAI configuration/delegation (CRUD) request. In some embodiments, CDTAIConReq includes the following parameters:

In some embodiments, the domain-specific AI trust engine is configured to configure corresponding (i.e., based on the desired domain-specific AIQoT) fairness and/or explainability and/or robustness methods/metrics/explanations in the domain-specific AI pipeline belonging to the cross-domain network service based on the cross-domain TAI configuration/delegation CRUD request. This configuration is illustrated in FIG9 by step 919. The implementation can be any suitable implementation, for example, such as described in PCT/EP2021/071044.

Depending on whether the configuration process in the previous step is successful, the domain-specific AI trust engine is configured to respond to the cross-domain AI trust engine with a cross-domain TAI fairness configuration/delegation CRUD response (CDTAIConResp). The response in some embodiments includes an ACK/NACK for satisfying the domain-specific AIQoT in the domain-specific AI pipeline belonging to the cross-domain network service. Generating and sending this response is shown in Figure 9 by step 921.

The operations shown as steps 923 , 925 , 927 , 929 and 931 in FIG. 9 illustrate a second alternative as part of a cross-domain TAI configuration API or a cross-domain TAI delegation API as shown as reference numeral 922 in FIG. 9 .

The cross-domain AI trust engine is configured to generate a cross-domain TAI configuration/delegation CRUD request (CDTAIConReq) and pass it to the domain-specific policy/intent manager 811 to notify the domain-specific policy/intent manager 811 of the converted domain-specific AIQoT that needs to be satisfied in the domain-specific AI pipeline belonging to the cross-domain network service. The operation of generating and passing CDTAIConReq is shown in Figure 9 by step 923. In some embodiments, CDTAIConReq may include the following parameters:

The domain-specific policy/intent manager 811 is configured to send the desired AIQoT information to the domain-specific AI trust engine 813, as shown by step 925 in Figure 9.

The domain-specific AI trust engine 813 can then convert the AIQoT into fairness, explainability and robustness requirements based on the cross-domain TAI configuration/delegated CRUD request, and configure the corresponding fairness and/or explainability and/or robustness methods/metrics/explanations in the domain-specific AI pipeline belonging to the cross-domain network service. This configuration is shown in Figure 9 by step 927. The implementation can be any suitable implementation, for example, such as described in PCT/EP2021/071044.

The domain-specific AI trust engine 813 may then be configured to send an ACK/NACK to the domain-specific policy/intent manager 811 for satisfying the desired AIQoT in the domain-specific AI pipeline belonging to the cross-domain network service. The operation of sending the ACK/NACK is shown in FIG. 9 by step 929.

Then, based on whether the domain-specific AIQoT is satisfied, the domain-specific policy/intent manager 811 is configured to respond to the cross-domain AI trust engine with a cross-domain TAI fairness configuration/delegation CRUD response (CDTAIConResp) containing ACK/NACK for satisfying the domain-specific AIQoT in the domain-specific AI pipeline belonging to the cross-domain network service. Generating and sending the cross-domain TAI fairness configuration/delegation CRUD response (CDTAIConResp) is shown in Figure 9 by step 931.

The operations shown in FIG. 9 as steps 933 , 935 , 937 , 939 and 941 illustrate an example cross-domain TAI reporting API or cross-domain TAI upgrading API implementation as shown in FIG. 9 as reference numeral 932 .

In some embodiments, the domain-specific AI trust engine 933 is configured to collect TAI reports (metrics and/or explanations) configured in step 919 or step 927 from the domain-specific AI pipeline belonging to the cross-domain network service. The collection of reports is illustrated in Figure 9 by step 933. The implementation can be any suitable implementation, for example, such as described in PCT/EP2021/071044.

In some embodiments, the network operator 701 generates and transmits a request or subscription for cross-domain TAI reporting, as shown by step 935 in FIG. 9 .

As shown in step 937 in FIG9 , a cross-domain TAI report request/subscription (CDTAIRReq/CDTAIRSub) is generated and sent from the cross-domain AI trust engine 801 to the domain-specific AI trust engine 813 of the report/subscription configuration of the domain-specific AI pipeline belonging to the cross-domain network service. In some embodiments, CDTAIRReq includes the following parameters:

In some embodiments, CDTAIRSub includes the following parameters:

In some embodiments, the cross-domain AI trust engine may store the cross-domain TAI report in a cross-domain trust knowledge database.

In some embodiments, when one or more reporting characteristics (i.e., periodic or on-demand) are met, the domain-specific AI trust engine 813 is configured to send a cross-domain TAI reporting response (TAIRResp) to the cross-domain AI trust engine 801 according to the reporting configuration specified in CDTAIRReq. In some embodiments, when the applicable TAI metric meets one or more reporting thresholds, the domain-specific AI trust engine 813 is configured to generate and send a cross-domain TAI reporting notification (TAIRNot) message to the cross-domain AI trust engine 801 including the actual TAI report. The generation of a cross-domain TAI reporting response (TAIRResp) or a cross-domain TAI reporting notification (TAIRNot) is shown in FIG. 9 by step 939.

The cross-domain AI trust engine 801 then performs root cause analysis of TAI reports from (multiple) domain-specific AI trust engines and provides a global view of issues with cross-domain network services (e.g., aggregated cross-domain network service-related TAI reports). For example, as shown by step 941 in Figure 9, cross-domain AIQoT violations can be sent to the network operator 701. The aggregation logic can consider combining/aggregating local/global interpretations of data/training/reasoning related received from (multiple) separate domain-specific AI trust engines.

10 shows a schematic diagram of non-volatile storage media 1000a (e.g., a computer disk (CD) or digital versatile disk (DVD)) and 1000b (e.g., a universal serial bus (USB) memory stick) storing instructions and/or parameters 1002 which, when executed by a processor, allow the processor to perform one or more steps of the method described above.

Note that while the above describes exemplifying embodiments, there are several variations and modifications which may be made to the disclosed solution without departing from the scope of the present invention.

It should be appreciated that although the above concepts have been discussed in the context of 5GS, one or more of these concepts may be applied to other cellular systems.

Therefore, the embodiments may vary within the scope of the appended claims. In general, some embodiments may be implemented in hardware or dedicated circuits, software, logic, or any combination thereof. For example, some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device, although the embodiments are not limited thereto. Although various embodiments may be illustrated and described as block diagrams, flow charts, or using some other graphical representations, it is well understood that, as non-limiting examples, the boxes, devices, systems, techniques, or methods described herein may be implemented in hardware, software, firmware, dedicated circuits or logic, general hardware or controllers or other computing devices, or some combination thereof.

These embodiments may be implemented by computer software stored in a memory and executable by at least one data processor of the entity involved, or by hardware, or by a combination of software and hardware. In addition, in this regard, it should be noted that any of the processes in, for example, Figures 7 and 8 may represent program steps, or interconnected logic circuits, boxes and functions, or a combination of program steps and logic circuits, boxes and functions. The software may be stored on physical media such as memory chips or memory blocks implemented within a processor, magnetic media (such as hard disks or floppy disks), and optical media (such as DVDs and their data variants CDs).

The memory may be of any type suitable for the local technical environment and may be implemented using any suitable data storage technology, such as semiconductor-based memory devices, magnetic memory devices and systems, optical memory devices and systems, fixed memory, and removable memory. The data processor may be of any type suitable for the local technical environment and may include, as non-limiting examples, one or more of the following: a general-purpose computer, a special-purpose computer, a microprocessor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a gate-level circuit, and a processor based on a multi-core processor architecture.

Alternatively or additionally, some embodiments may be implemented using a circuit system. The circuit system may be configured to perform one or more of the previously described functions and/or method steps. The circuit system may be provided in a base station and/or a communication device.

As used in this application, the term "circuitry" may refer to one or more or all of the following:

(a) hardware circuit implementation only (such as implementation only in analog and/or digital circuitry);

(b) a combination of hardware circuitry and software, such as:

(i) a combination of analog and/or digital hardware circuits and software/firmware, and

(ii) any portion of hardware processor(s) (including digital signal processor(s)) with software, software and memory(s) that work together to enable an apparatus (such as a communication device or base station) to perform the various functions previously described; and

(c) Hardware circuits and/or processor(s), such as microprocessor(s) or portions of microprocessor(s), that require software (e.g., firmware) to operate, but where the software is not required for operation, the software may not be present.

This definition of circuitry applies to all uses of the term in this application, including in any claims. As a further example, as used in this application, the term "circuitry" also covers an implementation of merely a hardware circuit or processor (or multiple processors) or a portion of a hardware circuit or processor and its (or their) accompanying software and/or firmware. The term circuitry also covers, for example, an integrated device.

The foregoing description has provided a complete and informative description of some embodiments by way of exemplary and non-limiting examples. However, various modifications and adaptations may become apparent to those skilled in the relevant art in view of the foregoing description when read in conjunction with the accompanying drawings and the appended claims. However, all such and similar modifications of the present teachings will still fall within the scope defined by the appended claims.

Claims (20)

1. An apparatus, comprising components configured to facilitate trustworthiness functions of a cross-domain network service-related machine learning or artificial intelligence pipeline, wherein the cross-domain machine learning or artificial intelligence pipeline is used to control a network including at least two domains Cognitive autonomous networks. 2. The apparatus of claim 1, wherein the component configured to facilitate cross-domain network service-related machine learning or artificial intelligence pipeline trustworthiness functions is configured to facilitate at least one of the following, wherein the cross-domain Machine learning or artificial intelligence pipelines are used to control cognitive autonomous networks that include at least two domains: Define the trustworthiness of machine learning or artificial intelligence pipelines related to cross-domain network services; Configure the trustworthiness of machine learning or artificial intelligence pipelines related to cross-domain network services; Measuring the trustworthiness of machine learning or artificial intelligence pipelines associated with cross-domain network services; and Reports on the trustworthiness of machine learning or artificial intelligence pipelines related to cross-domain network services. 3. The device according to any one of claims 1 or 2, wherein the cross-domain network service-related machine learning or artificial intelligence pipeline credibility function includes at least one of the following: fairness; Explainability; and robustness. 4. The apparatus of any one of claims 1 to 3, wherein the component configured to facilitate cross-domain network service-related machine learning or artificial intelligence pipeline trustworthiness functions is configured for, Wherein the cross-domain machine learning or artificial intelligence pipeline is used to control a cognitive autonomous network including at least two domains: Obtain cross-domain machine learning or artificial intelligence quality credibility, and the cross-domain machine learning or artificial intelligence quality credibility is configured to: cover domain-specific machine learning or artificial intelligence quality credibility requirements, and cross-domain network services Constraints on relevant machine learning or artificial intelligence pipelines; converting the cross-domain machine learning or artificial intelligence quality confidence into at least one domain-specific machine learning or artificial intelligence quality confidence for at least one of the at least two domains; and Confidence in the at least one domain-specific machine learning or artificial intelligence quality is provided for at least one of the at least two domains. 5. The apparatus according to claim 4, wherein the component configured to convert the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality for at least one of the at least two domains is configured to: convert the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality based on the risk level of the cross-domain network service. 6. An apparatus according to any one of claims 4 or 5, wherein the component configured to convert the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality for at least one of the at least two domains is configured to: based on at least one service level agreement requirement for the cross-domain network, convert the cross-domain machine learning or artificial intelligence credibility quality into at least one domain-specific machine learning or artificial intelligence credibility quality, wherein the at least one service level agreement includes at least one of the following: service type; service priority; and at least one key performance indicator metric. 7. The apparatus of any one of claims 4 to 6, configured to provide the at least one domain-specific machine learning or artificial intelligence trust for at least one of the at least two domains. The components of degree quality are configured to: Generate a cross-domain credibility machine learning or artificial intelligence configuration or delegation request and pass it to at least one domain-specific machine learning or artificial intelligence credibility function, the cross-domain credibility machine learning or artificial intelligence configuration or The delegate request is configured to: control the at least one domain-specific machine learning or artificial intelligence trustworthiness functionality to configure a machine learning or artificial intelligence pipeline for the at least one of the at least two domains; and Obtained from the at least one domain-specific machine learning or artificial intelligence credibility function based on implementation of the configuration of the machine learning or artificial intelligence pipeline for the at least one of the at least two domains Configure or delegate responses with cross-domain trustworthiness machine learning or artificial intelligence. 8. The apparatus of any one of claims 4 to 6, configured to provide the at least one domain-specific machine learning or artificial intelligence trust for at least one of the at least two domains. Quality components are configured as: generating and passing a cross-domain trustworthy machine learning or artificial intelligence configuration or delegation request to at least one domain-specific policy manager, the cross-domain trustworthy machine learning or artificial intelligence configuration or delegation request being configured to: control at least one domain-specific machine learning or artificial intelligence trustworthy function to configure a machine learning or artificial intelligence pipeline for at least one of the at least two domains; obtaining cross-domain credibility machine learning from the at least one domain policy manager based on implementation of the configuration of the machine learning or artificial intelligence pipeline for the at least one of the at least two domains; AI configuration or delegate response. 9. The apparatus according to any one of claims 7 or 8, wherein the cross-domain credibility machine learning or artificial intelligence configuration or delegation request includes: a domain scope parameter configured to identify the domain being addressed by the request; a pipeline identification parameter configured to identify the domain-specific machine learning or artificial intelligence pipeline that the request is addressing; A category identification parameter configured to identify the at least one domain-specific machine learning or artificial intelligence credibility quality. 10. The apparatus of claim 9, when dependent on claim 7, wherein the cross-domain credibility machine learning or artificial intelligence configuration or delegation request further includes at least one of the following: a desired fairness parameter configured to indicate a relative fairness level for the domain-specific machine learning or artificial intelligence pipeline; a desired explainability parameter configured to indicate a desired level of explainability for the domain-specific machine learning or artificial intelligence pipeline; a desired technical robustness parameter configured to indicate a desired technical robustness level for the domain-specific machine learning or artificial intelligence pipeline; and A desired adversarial robustness parameter configured to indicate a desired level of adversarial robustness for the domain-specific machine learning or artificial intelligence pipeline. 11. The apparatus according to any one of claims 4 to 10, wherein the component configured to facilitate cross-domain network service-related machine learning or artificial intelligence pipeline credibility functions is configured for the following items, wherein the cross-domain machine learning or artificial intelligence pipeline is used to control a cognitive autonomous network including at least two domains: Generate a cross-domain credibility machine learning or artificial intelligence capability information request and pass it to at least one domain-specific machine learning or artificial intelligence credibility function, the cross-domain credibility machine learning or artificial intelligence capability information request configured to: control the at least one domain-specific machine learning or artificial intelligence credibility function to enable capability discovery of the machine learning or artificial intelligence pipeline for the at least one of the at least two domains; as well as Obtain a cross-domain credibility machine learning or artificial intelligence capability information response from the at least one domain-specific machine learning or artificial intelligence credibility function, the artificial intelligence capability information response report is for use in the at least two domains The capability discovery of the machine learning or artificial intelligence pipeline of the at least one domain. 12. The apparatus according to claim 11, wherein the cross-domain credibility machine learning or artificial intelligence capability information request includes: a domain scope parameter configured to identify the domain to which the request is being addressed; and A scope parameter configured to identify the domain-specific machine learning or artificial intelligence pipeline that the request is addressing. 13. The apparatus of claim 12, wherein the cross-domain credibility machine learning or artificial intelligence capability information request further includes a pipeline stage parameter configured to identify all of the requests being addressed. Describe the stages of a domain-specific machine learning or artificial intelligence pipeline. 14. The apparatus of any one of claims 4 to 13, wherein the component configured to facilitate cross-domain network service-related machine learning or artificial intelligence pipeline trustworthiness functions is further configured to: Wherein the cross-domain machine learning or artificial intelligence pipeline is used to control a cognitive autonomous network including at least two domains: Obtain cross-domain credibility artificial intelligence report requests or subscriptions from network operators; Generate a domain-specific trustworthiness AI report request or subscription based on the cross-domain trustworthiness AI report request or subscription from the network operator and pass it to at least one domain-specific machine learning or artificial intelligence pipeline a trustworthiness function, wherein the domain-specific trustworthiness artificial intelligence report request or subscription is configured to: control the at least one domain-specific machine learning or artificial intelligence pipeline trustworthiness function to provide at least one domain-specific Machine learning or artificial intelligence pipeline reporting responses or notifications; Receive machine learning or artificial intelligence capability information from the at least one domain-specific machine learning or artificial intelligence pipeline credibility function, and/or the at least one domain-specific machine learning or artificial intelligence pipeline reporting response or notification; Store machine learning or artificial intelligence capability information received from the at least one domain-specific machine learning or artificial intelligence pipeline, and/or the at least one domain-specific machine learning or artificial intelligence pipeline report in a cross-domain trust knowledge database ; The cross-domain credibility artificial intelligence report is generated and delivered based on the at least one domain-specific machine learning or artificial intelligence pipeline report response or notification. 15. The apparatus of claim 14, wherein the cross-domain credibility artificial intelligence report request includes: a domain scope parameter configured to identify the domain to which the request is being addressed; a scope parameter configured to identify the domain-specific machine learning or artificial intelligence pipeline to which the request is addressed; and A pipeline stage parameter configured to identify the stage of the domain-specific machine learning or artificial intelligence pipeline that the request is addressing. 16. The apparatus of claim 15, wherein the cross-domain credibility artificial intelligence report request further includes at least one of the following: A list of fairness metric parameters configured to identify the fairness metrics to be reported; A list of fairness metric interpretations configured to identify the fairness metric interpretations to be reported; A list of explainability metrics configured to identify the explainability metrics to be reported; A list of interpretations, configured to identify interpretations to be reported; A list of technical robustness metrics configured to identify technical robustness metrics to be reported; A list of technical robustness metric interpretations configured to identify technical robustness metric interpretations to be reported; A list of adversarial robustness metrics configured to identify adversarial robustness metrics to be reported; A list of adversarial robustness metric interpretations configured to identify adversarial robustness metric interpretations to be reported; A start time parameter, configured to identify the start time for reporting; An end time parameter configured to identify the end time for reporting; and Report interval parameter, configured to identify the periodic interval used for reporting. 17. The apparatus of claim 14, wherein the cross-domain credibility artificial intelligence reporting subscription includes: a domain scope parameter configured to identify the domain being addressed by the subscription; A scope parameter configured to identify the domain-specific machine learning or artificial intelligence pipeline that the subscription is addressing; and A pipeline stage parameter configured to identify the stage of the domain-specific machine learning or artificial intelligence pipeline that the subscription is addressing. 18. The apparatus of claim 17, wherein the cross-domain credibility artificial intelligence report subscription further includes at least one of the following: A list of fairness metric parameters configured to identify the fairness metrics to be reported; A list of explainability metrics configured to identify the explainability metrics to be reported; a list of technical robustness metrics configured to identify technical robustness metrics to be reported; a list of adversarial robustness metrics configured to identify adversarial robustness metrics to be reported; and Cross reporting threshold parameter, configured to identify the threshold for which a metric or metric interpretation is reported. 19. A method comprising: Facilitate cross-domain network service-related machine learning or artificial intelligence pipeline trustworthiness functions, wherein the cross-domain machine learning or artificial intelligence pipeline is used to control a cognitive autonomous network including at least two domains. 20. A computer program comprising computer-executable instructions which, when run on one or more processors, perform the steps of the method of claim 19.