[go: up one dir, main page]

MX2018011305A - Técnicas para corregir el desvío de entrenamiento lingüístico en los datos de entrenamiento. - Google Patents

Técnicas para corregir el desvío de entrenamiento lingüístico en los datos de entrenamiento.

Info

Publication number
MX2018011305A
MX2018011305A MX2018011305A MX2018011305A MX2018011305A MX 2018011305 A MX2018011305 A MX 2018011305A MX 2018011305 A MX2018011305 A MX 2018011305A MX 2018011305 A MX2018011305 A MX 2018011305A MX 2018011305 A MX2018011305 A MX 2018011305A
Authority
MX
Mexico
Prior art keywords
model
training
training data
sentences
language
Prior art date
Application number
MX2018011305A
Other languages
English (en)
Inventor
Shroff Gautam
Vig Lovekesh
Agarwal Puneet
Patidar Mayur
Original Assignee
Tata Consultancy Services Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tata Consultancy Services Ltd filed Critical Tata Consultancy Services Ltd
Publication of MX2018011305A publication Critical patent/MX2018011305A/es

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/332—Query formulation
    • G06F16/3329—Natural language query formulation
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/245—Query processing
    • G06F16/2455—Query execution
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G06N3/0455—Auto-encoder networks; Encoder-decoder networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0475—Generative networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/09—Supervised learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/04—Inference or reasoning models
    • G06N5/046—Forward inferencing; Production systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Probability & Statistics with Applications (AREA)
  • Human Computer Interaction (AREA)
  • Machine Translation (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

En los sistemas de asistente automatizado, se utiliza un modelo de aprendizaje profundo en forma de un clasificador de memoria larga a corto plazo (LSTM) para asignar preguntas a clases, con cada clase teniendo una respuesta depurada manualmente. Un equipo de expertos crea manualmente los datos de entrenamiento utilizados para entrenar a este clasificador. Confiar en la depuración humana a menudo hace que dichos desvíos de entrenamiento lingüístico se arraiguen en los datos de entrenamiento, ya que cada individuo tiene un estilo específico de escribir el lenguaje natural y usa algunas palabras en un contexto específico solamente. Los modelos profundos terminan aprendiendo estos desvíos, en lugar de las palabras conceptuales centrales de las clases objetivo. Para corregir estos desvíos, las oraciones significativas se generan automáticamente usando un modelo generativo, y luego se usan para entrenar un modelo de clasificación. Por ejemplo, se utiliza un auto-codificador variacional (VAE) como modelo generativo para generar oraciones novedosas y se utiliza un modelo de lenguaje (LM) para seleccionar oraciones basadas en la probabilidad.
MX2018011305A 2017-09-18 2018-09-17 Técnicas para corregir el desvío de entrenamiento lingüístico en los datos de entrenamiento. MX2018011305A (es)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
IN201721033035 2017-09-18

Publications (1)

Publication Number Publication Date
MX2018011305A true MX2018011305A (es) 2019-07-04

Family

ID=65721524

Family Applications (1)

Application Number Title Priority Date Filing Date
MX2018011305A MX2018011305A (es) 2017-09-18 2018-09-17 Técnicas para corregir el desvío de entrenamiento lingüístico en los datos de entrenamiento.

Country Status (6)

Country Link
US (1) US11373090B2 (es)
JP (1) JP6606243B2 (es)
AU (1) AU2018232914B2 (es)
BR (1) BR102018068925A2 (es)
CA (1) CA3017655C (es)
MX (1) MX2018011305A (es)

Families Citing this family (32)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107832439B (zh) * 2017-11-16 2019-03-08 百度在线网络技术（北京）有限公司 多轮状态追踪的方法、系统及终端设备
US11270082B2 (en) 2018-08-20 2022-03-08 Verint Americas Inc. Hybrid natural language understanding
US11087184B2 (en) * 2018-09-25 2021-08-10 Nec Corporation Network reparameterization for new class categorization
US10909671B2 (en) * 2018-10-02 2021-02-02 International Business Machines Corporation Region of interest weighted anomaly detection
US11217226B2 (en) * 2018-10-30 2022-01-04 Verint Americas Inc. System to detect and reduce understanding bias in intelligent virtual assistants
US10963645B2 (en) * 2019-02-07 2021-03-30 Sap Se Bi-directional contextualized text description
US11003861B2 (en) 2019-02-13 2021-05-11 Sap Se Contextualized text description
US11604927B2 (en) 2019-03-07 2023-03-14 Verint Americas Inc. System and method for adapting sentiment analysis to user profiles to reduce bias
US11922301B2 (en) * 2019-04-05 2024-03-05 Samsung Display Co., Ltd. System and method for data augmentation for trace dataset
CN110090016B (zh) * 2019-04-28 2021-06-25 心医国际数字医疗系统(大连)有限公司 定位r波位置的方法及系统、使用lstm神经网络的r波自动检测方法
US20200380309A1 (en) * 2019-05-28 2020-12-03 Microsoft Technology Licensing, Llc Method and System of Correcting Data Imbalance in a Dataset Used in Machine-Learning
CN110297886A (zh) * 2019-05-31 2019-10-01 广州大学 基于短文本的oj题目分类器构建方法及题目模拟方法
WO2020247586A1 (en) 2019-06-06 2020-12-10 Verint Americas Inc. Automated conversation review to surface virtual assistant misunderstandings
US10937213B2 (en) * 2019-06-27 2021-03-02 Fuji Xerox Co., Ltd. Systems and methods for summarizing and steering multi-user collaborative data analyses
CN110647627B (zh) * 2019-08-06 2022-05-27 北京百度网讯科技有限公司 答案生成方法及装置、计算机设备与可读介质
CN110580289B (zh) * 2019-08-28 2021-10-29 浙江工业大学 一种基于堆叠自动编码器和引文网络的科技论文分类方法
CN110795913B (zh) * 2019-09-30 2024-04-12 北京大米科技有限公司 一种文本编码方法、装置、存储介质及终端
US11710045B2 (en) 2019-10-01 2023-07-25 Samsung Display Co., Ltd. System and method for knowledge distillation
CN110941964B (zh) 2019-12-11 2023-08-15 北京小米移动软件有限公司 双语语料筛选方法、装置及存储介质
US11270080B2 (en) 2020-01-15 2022-03-08 International Business Machines Corporation Unintended bias detection in conversational agent platforms with machine learning model
US11610079B2 (en) * 2020-01-31 2023-03-21 Salesforce.Com, Inc. Test suite for different kinds of biases in data
US11631036B2 (en) 2020-05-11 2023-04-18 Fujitsu Limited Bias mitigation in machine learning pipeline
CN111624606B (zh) * 2020-05-27 2022-06-21 哈尔滨工程大学 一种雷达图像降雨识别方法
CN111738364B (zh) * 2020-08-05 2021-05-25 国网江西省电力有限公司供电服务管理中心 一种基于用户负荷与用电参量相结合的窃电检测方法
KR20220118123A (ko) * 2021-02-18 2022-08-25 현대자동차주식회사 질의응답 시스템 및 그 제어 방법
US11947908B2 (en) * 2021-04-07 2024-04-02 Baidu Usa Llc Word embedding with disentangling prior
CN113204641B (zh) * 2021-04-12 2022-09-02 武汉大学 一种基于用户特征的退火注意力谣言鉴别方法及装置
US12374321B2 (en) * 2021-06-08 2025-07-29 Microsoft Technology Licensing, Llc Reducing biases of generative language models
CN113535549B (zh) * 2021-06-22 2024-08-20 科大讯飞股份有限公司 测试数据的扩充方法、装置、设备及计算机可读存储介质
KR102729294B1 (ko) * 2021-11-01 2024-11-11 부산대학교 산학협력단 실시간 배치셋 기반으로 인공지능 학습 데이터셋의 공정성을 보정하는 장치 및 그 방법
US12032564B1 (en) * 2023-01-03 2024-07-09 Sap Se Transforming natural language request into enterprise analytics query using fine-tuned machine learning model
CN120077370A (zh) * 2023-06-01 2025-05-30 谷歌有限责任公司 使用语言模型从查询中生成检索词元

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7603330B2 (en) * 2006-02-01 2009-10-13 Honda Motor Co., Ltd. Meta learning for question classification
US20140214401A1 (en) 2013-01-29 2014-07-31 Tencent Technology (Shenzhen) Company Limited Method and device for error correction model training and text error correction
US10909329B2 (en) * 2015-05-21 2021-02-02 Baidu Usa Llc Multilingual image question answering
JP6618735B2 (ja) * 2015-08-31 2019-12-11 国立研究開発法人情報通信研究機構 質問応答システムの訓練装置及びそのためのコンピュータプログラム
US20180357531A1 (en) 2015-11-27 2018-12-13 Devanathan GIRIDHARI Method for Text Classification and Feature Selection Using Class Vectors and the System Thereof
CN109792402B (zh) * 2016-07-08 2020-03-06 艾赛普公司 自动响应用户的请求
WO2019000326A1 (en) * 2017-06-29 2019-01-03 Microsoft Technology Licensing, Llc Generating responses in automated chatting

Also Published As

Publication number Publication date
BR102018068925A2 (pt) 2019-05-28
CA3017655C (en) 2021-04-20
CA3017655A1 (en) 2019-03-18
AU2018232914B2 (en) 2020-07-02
JP2019057280A (ja) 2019-04-11
JP6606243B2 (ja) 2019-11-13
US20190087728A1 (en) 2019-03-21
AU2018232914A1 (en) 2019-04-04
US11373090B2 (en) 2022-06-28

Similar Documents

Publication Publication Date Title
MX2018011305A (es) Técnicas para corregir el desvío de entrenamiento lingüístico en los datos de entrenamiento.
CL2015002614A1 (es) Predicción de texto basados en múltiples modelos de lenguaje.
Chandiok et al. CIT: Integrated cognitive computing and cognitive agent technologies based cognitive architecture for human-like functionality in artificial systems
Batrova et al. Bilingual teaching high school students of information communication technologies by means of the Russian and English languages at schools of Tatarstan
Hapsari et al. The Application of Riddle Game in Teaching Speaking for the Eighth Grade Students of MTsN Kedunggalar Ngawi in the Schooling Year 2014/2015
González-Flores et al. Here to Stay After the Pandemic: Sentiment Analysis of Higher Education Teachers at UNAM
Lozano et al. Linking education for sustainable development and corporate sustainability: developing an organisational change management for corporate sustainability course
Anna Architecture of modern educational environment: a global evolutionary approach
Andino Jaramillo Teacher training: Pillar for the identification and bullying management
Apte Talent in Transition: The Dynamics and Constants of Transience
Yuliansyah et al. Teaching Vocabulary to Junior High School Students by Using Kick Me Technique
Toala-Sanchez et al. Cognitive Abilities and Performance in Database Design: An Observational Study
Burton Book Review: returns: becoming indigenous in the twenty-first century by James Clifford
Rahmasari Using Newspapers as Learning Media to Teach Reading Skills
ANDEWI THE STUDENTS’PERCEPTIONS OF RECIPROCAL TEACHING: STUDENTS’ANXIETY AND SELF-EFFICACY IN READING
Sujatmoko et al. The values of local wisdom in Dedeng: Avocal tradition of Malay Langkatcommunity
Siliezar RWU and Grupo Puntacana to Partner in Building Ornamental Fish Hatchery in Dominican Republic
Woolson Teaching Ecocriticism and the Global South
SA119410256B1 (ar) نظام للتعليم عن بعد قائم على التواصل غير المتزامن
Melinescu THE IMPROVEMENT OF EDUCATIONAL COMMUNICATION THROUGH ESPERE METHOD
Knowles Aubrey Blanche Atlassian global head of diversity
Drljača Margić et al. Language in EMI as a motivating factor, perceived barrier and essential prerequisite
Biasin et al. Adult Education Policies in Italy
Delgado WOMEN ENTREPRENEURSHIP USING THE “WORKING WITH PEOPLE” MODEL: THIS IS A STUDY CASE FROM THE “TEJEMUJERES” BUSINESS CENTER FOR WOMEN, IN THE GUALACEO TOWN, AZUAY PROVINCE OF ECUADOR.
Maluleke A study of the academic writing proficiency of level one students at the University of Venda