KR102684665B1 - Method for training diffusion model based on the type of data - Google Patents

Abstract

According to one embodiment of the present disclosure, disclosed is a method for training a neural network model for generating data, which is performed by one or more processors of a computing device. The method may comprise the steps of: receiving original data and determining a type of the original data; determining a noise ratio based on the determined type of the original data; adding noise to the original data based on the determined noise ratio to obtain first data; predicting noise included in the first data using the neural network model; comparing the predicted noise and the noise added to the original data to calculate a loss function; and training the neural network model based on the loss function.

Description

A method for learning a diffusion model based on the type of data {METHOD FOR TRAINING DIFFUSION MODEL BASED ON THE TYPE OF DATA}

This disclosure relates to a method for learning a diffusion model based on the type of data. More specifically, in the process of learning the diffusion model to generate data, the diffusion model is learned by setting the noise ratio differently depending on the type of data. It's about how to do it.

The noise schedule methods of the existing widely known diffusion models were optimized for generating image data with 8-bit depth data per channel. In other words, since image data has a depth of 8 bits per channel, noise components smaller than 1/(2^8) (the minimum size of data that can express an image) contained in the generated result cannot be ignored. However, in the case of audio data, each sample has a depth of 16 bits, so the result generated through the diffusion model has a size of less than 1/(2^8) and 1/(2^16). If noise components larger than the size were included, there was a problem that could not be ignored. Therefore, when generating speech using a diffusion model learned by applying the noise schedule used in the image domain, there is a problem that audible noise may be generated. did.

Therefore, by learning a diffusion model to generate data by applying a different noise schedule depending on the type of data, the diffusion model can be specialized in generating the desired type of data. The need for methods is emerging.

Meanwhile, the present disclosure has been derived based at least on the technical background examined above, but the technical problem or purpose of the present disclosure is not limited to solving the problems or shortcomings examined above. In other words, in addition to the technical issues discussed above, the present disclosure can cover various technical issues related to the content to be described below.

The present disclosure aims to minimize the noise contained in data generated of a desired type by learning a diffusion model to generate data by applying a different noise schedule depending on the type of data. do.

Meanwhile, the technical problem to be achieved by the present disclosure is not limited to the technical problems mentioned above, and may include various technical problems within the scope of what is apparent to those skilled in the art from the contents described below.

A method performed by a computing device according to an embodiment of the present disclosure for realizing the above-described problem is disclosed. The method includes receiving original data and determining a type of the original data; determining a noise ratio based on the determined type of original data; acquiring first data by adding noise to the original data based on the determined noise ratio; predicting noise included in the first data using a neural network model; Comparing the predicted noise and noise added to the original data to calculate a loss function; And it may include training the neural network model based on the loss function.

Alternatively, the neural network model may include a model for removing part or all of the predicted noise with respect to Gaussian distributed noise, thereby obtaining data from which part or all of the predicted noise has been removed. You can.

Alternatively, determining the ratio of noise based on the determined type of original data may include determining the ratio of noise based on the size of data that can be represented for the determined type of original data. there is.

Alternatively, the step of determining the ratio of noise based on the size of data that can be represented for the determined type of original data includes adding more noise to the original data than the minimum size of data that can be represented for the determined type of original data. It may include determining the ratio of the noise so that the size of the noise is small.

Alternatively, the type of original data may include image data or audio data.

Alternatively, acquiring first data by adding noise to the original data based on the determined ratio of noise may include adding first noise to the original data based on the determined ratio of noise to obtain first data. Obtaining a; and acquiring second data by adding second noise to the first data based on the determined noise ratio.

Alternatively, predicting noise included in the first data using the neural network model may include predicting first noise included in the first data using the neural network model; and predicting second noise included in the second data using the neural network model, wherein calculating a loss function by comparing the predicted noise and the noise added to the original data includes: Comparing the predicted first noise and the first noise added to the original data to calculate a first loss function; and calculating a second loss function by comparing the predicted second noise and the second noise added to the first data, wherein the step of training the neural network model based on the loss function includes: It may include training the neural network model based on at least one of the first loss function or the second loss function.

According to an embodiment of the present disclosure for realizing the above-described object, a computer program stored in a computer-readable storage medium is disclosed. The computer program, when executed on one or more processors, causes the one or more processors to perform operations for training a neural network model for generating data, the operations comprising: receiving original data, and types of the original data. The action of determining; determining a noise ratio based on the determined type of original data; acquiring first data by adding noise to the original data based on the determined noise ratio; Predicting noise included in the first data using a neural network model; calculating a loss function by comparing the predicted noise and noise added to the original data; and training the neural network model based on the loss function.

Alternatively, the operation of determining the ratio of noise based on the determined type of original data may include determining the ratio of noise based on the size of data that can be expressed for the determined type of original data. there is.

Alternatively, the operation of determining the ratio of the noise based on the size of data representable for the determined type of original data adds to the original data more than the minimum size of data representable for the determined type of original data. It may include an operation of determining the ratio of the noise so that the size of the noise is small.

Alternatively, the operation of acquiring first data by adding noise to the original data based on the determined ratio of noise may include adding first noise to the original data based on the determined ratio of noise to obtain first data. An operation to obtain; and acquiring second data by adding second noise to the first data based on the determined noise ratio.

Alternatively, predicting noise included in the first data using the neural network model may include predicting first noise included in the first data using the neural network model; and predicting second noise included in the second data using the neural network model, and calculating a loss function by comparing the predicted noise with the noise added to the original data. Comparing the predicted first noise and the first noise added to the original data to calculate a first loss function; and an operation of calculating a second loss function by comparing the predicted second noise and the second noise added to the first data, and the operation of training the neural network model based on the loss function includes: It may include training the neural network model based on at least one of the first loss function or the second loss function.

A computing device according to an embodiment of the present disclosure for realizing the above-described problem is disclosed. The device includes at least one processor; and a memory, wherein the at least one processor receives original data and determines a type of the original data; determine a noise ratio based on the determined type of original data; Obtaining first data by adding noise to the original data based on the determined noise ratio; predict noise included in the first data using a neural network model; Compare the predicted noise and the noise added to the original data to calculate a loss function; And it may be configured to learn the neural network model based on the loss function.

Alternatively, the process of determining the ratio of noise based on the determined type of original data may include determining the ratio of noise based on the size of data that can be expressed for the determined type of original data. there is.

Alternatively, the process of determining the ratio of the noise based on the size of data that can be represented for the determined type of original data includes adding more noise to the original data than the minimum size of data that can be represented for the determined type of original data. It may include a process of determining the ratio of the noise so that the size of the noise is small.

Alternatively, the process of acquiring first data by adding noise to the original data based on the determined ratio of noise may include adding first noise to the original data based on the determined ratio of noise to obtain first data. obtain; And it may include obtaining second data by adding second noise to the first data based on the determined noise ratio.

Alternatively, the process of predicting noise included in the first data using the neural network model may include predicting first noise included in the first data using the neural network model; And a process of predicting second noise included in the second data using the neural network model, and calculating a loss function by comparing the predicted noise and the noise added to the original data, calculate a first loss function by comparing the predicted first noise and the first noise added to the original data; And a process of calculating a second loss function by comparing the predicted second noise and the second noise added to the first data, and the process of learning the neural network model based on the loss function includes: It may include training the neural network model based on at least one of the first loss function or the second loss function.

A data structure included in a computer-readable storage medium according to an embodiment of the present disclosure for realizing the above-described problem is disclosed. The data structure corresponds to parameters of a neural network, at least in part updated during a learning process, and operation of the neural network is based at least in part on the parameters, the learning process comprising: receiving original data, determining the type; determining a noise ratio based on the determined type of original data; acquiring first data by adding noise to the original data based on the determined noise ratio; predicting noise included in the first data using a neural network model; Comparing the predicted noise and noise added to the original data to calculate a loss function; And it may include training the neural network model based on the loss function.

A method performed by a computing device according to an embodiment of the present disclosure for realizing the above-described problem is disclosed. The method includes determining the type of data to be generated; And using a neural network model, removing part or all of the noise for Gaussian distributed noise, and obtaining the determined type of data from which part or all of the noise has been removed, the neural network The model may be a model learned by considering a loss function associated with a noise ratio determined according to the type of data.

The present disclosure learns a diffusion model to generate data by applying a different noise schedule depending on the type of data, thereby minimizing noise included in data generated of the desired type and producing better quality. You can achieve the effect of generating data.

Meanwhile, the effects of the present disclosure are not limited to the effects mentioned above, and various effects may be included within the range apparent to those skilled in the art from the contents described below.

1 is a block diagram of a computing device for learning a diffusion model based on the type of data according to an embodiment of the present disclosure.
Figure 2 is a schematic diagram showing a network function according to an embodiment of the present disclosure.
Figure 3 is a flowchart showing a method for learning a neural network model based on the type of data according to an embodiment of the present disclosure.
Figure 4 is a neural network model to obtain data from which part or all of the predicted noise is removed by removing part or all of the predicted noise for Gaussian distributed noise according to an embodiment of the present disclosure. This is a schematic diagram to explain the learning process.
Figure 5 is a schematic diagram illustrating a process of determining the type of received original data and determining the noise ratio based on the determined type of original data according to an embodiment of the present disclosure.
6 illustrates determining the type of data to be generated according to an embodiment of the present disclosure, and obtaining the determined type of data by utilizing a neural network model learned by considering the loss function associated with the ratio of noise determined according to the type of data. This is a flowchart showing the process.
7 is a brief, general schematic diagram of an example computing environment in which embodiments of the present disclosure may be implemented.

Various embodiments are now described with reference to the drawings. In this specification, various descriptions are presented to provide an understanding of the disclosure. However, it is clear that these embodiments may be practiced without these specific descriptions.

As used herein, the terms “component,” “module,” “system,” and the like refer to a computer-related entity, hardware, firmware, software, a combination of software and hardware, or an implementation of software. For example, a component may be, but is not limited to, a process running on a processor, a processor, an object, a thread of execution, a program, and/or a computer. For example, both an application running on a computing device and the computing device can be a component. One or more components may reside within a processor and/or thread of execution. A component may be localized within one computer. A component may be distributed between two or more computers. Additionally, these components can execute from various computer-readable media having various data structures stored thereon. Components may transmit signals, for example, with one or more data packets (e.g., data and/or signals from one component interacting with other components in a local system, a distributed system, to other systems and over a network such as the Internet). Depending on the data being transmitted, they may communicate through local and/or remote processes.

Additionally, the term “or” is intended to mean an inclusive “or” and not an exclusive “or.” That is, unless otherwise specified or clear from context, “X utilizes A or B” is intended to mean one of the natural implicit substitutions. That is, either X uses A; X uses B; Or, if X uses both A and B, “X uses A or B” can apply to either of these cases. Additionally, the term “and/or” as used herein should be understood to refer to and include all possible combinations of one or more of the related listed items.

Additionally, the terms “comprise” and/or “comprising” should be understood to mean that the corresponding feature and/or element is present. However, the terms “comprise” and/or “comprising” should be understood as not excluding the presence or addition of one or more other features, elements and/or groups thereof. Additionally, unless otherwise specified or the context is clear to indicate a singular form, the singular terms herein and in the claims should generally be construed to mean “one or more.”

And, the term “at least one of A or B” should be interpreted to mean “a case containing only A,” “a case containing only B,” and “a case of combining A and B.”

Those skilled in the art will additionally understand that the various illustrative logical blocks, configurations, modules, circuits, means, logic, and algorithm steps described in connection with the embodiments disclosed herein may be implemented in electronic hardware, computer software, or both. It should be recognized that it can be implemented with combinations of. To clearly illustrate the interchangeability of hardware and software, various illustrative components, blocks, configurations, means, logics, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented in hardware or software will depend on the specific application and design constraints imposed on the overall system. A skilled technician can implement the described functionality in a variety of ways for each specific application. However, such implementation decisions should not be construed as causing a departure from the scope of the present disclosure.

The description of the presented embodiments is provided to enable anyone skilled in the art to use or practice the present invention. Various modifications to these embodiments will be apparent to those skilled in the art. The general principles defined herein may be applied to other embodiments without departing from the scope of the disclosure. Therefore, the present invention is not limited to the embodiments presented herein. The present invention is to be interpreted in the broadest scope consistent with the principles and novel features presented herein.

In this disclosure, network function, artificial neural network, and neural network may be used interchangeably.

1 is a block diagram of a computing device for learning a diffusion model based on the type of data according to an embodiment of the present disclosure.

The configuration of the computing device 100 shown in FIG. 1 is only a simplified example. In one embodiment of the present disclosure, the computing device 100 may include different components for performing the computing environment of the computing device 100, and only some of the disclosed components may configure the computing device 100.

The computing device 100 may include a processor 110, a memory 130, and a network unit 150.

The processor 110 may be composed of one or more cores, and may include a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), and a tensor processing unit (TPU) of a computing device. unit) may include a processor for data analysis and deep learning. The processor 110 may read a computer program stored in the memory 130 and perform data processing for machine learning according to an embodiment of the present disclosure. According to an embodiment of the present disclosure, the processor 110 may perform an operation for learning a neural network model. The processor 110 processes input data for learning in deep learning (DL), extracts features from input data, calculates errors, and learns neural network models such as updating the weights of the neural network model using backpropagation. You can perform calculations for . At least one of the CPU, GPGPU, and TPU of the processor 110 may process learning of the neural network model. For example, the CPU and GPGPU can work together to process neural network model learning and data classification using the neural network model. Additionally, in one embodiment of the present disclosure, the processors of a plurality of computing devices can be used together to process learning of a neural network model and data classification using the neural network model. Additionally, a computer program executed in a computing device according to an embodiment of the present disclosure may be a CPU, GPGPU, or TPU executable program.

According to an embodiment of the present disclosure, the memory 130 may store any type of information generated or determined by the processor 110 and any type of information received by the network unit 150.

According to an embodiment of the present disclosure, the memory 130 is a flash memory type, hard disk type, multimedia card micro type, or card type memory (e.g. (e.g. SD or -Only Memory), and may include at least one type of storage medium among magnetic memory, magnetic disk, and optical disk. The computing device 100 may operate in connection with web storage that performs a storage function of the memory 130 on the Internet. The description of the memory described above is merely an example, and the present disclosure is not limited thereto.

The network unit 150 according to an embodiment of the present disclosure includes Public Switched Telephone Network (PSTN), x Digital Subscriber Line (xDSL), Rate Adaptive DSL (RADSL), Multi Rate DSL (MDSL), and VDSL ( A variety of wired communication systems can be used, such as Very High Speed DSL), Universal Asymmetric DSL (UADSL), High Bit Rate DSL (HDSL), and Local Area Network (LAN).

In addition, the network unit 150 presented in the present disclosure includes Code Division Multi Access (CDMA), Time Division Multi Access (TDMA), Frequency Division Multi Access (FDMA), Orthogonal Frequency Division Multi Access (OFDMA), and SC-FDMA ( A variety of wireless communication systems can be used, such as Single Carrier-FDMA) and other systems.

In the present disclosure, the network unit 150 may be configured regardless of the communication mode, such as wired or wireless, and may be composed of various communication networks such as a personal area network (PAN) and a wide area network (WAN). You can. Additionally, the network may be the well-known World Wide Web (WWW), and may also use wireless transmission technology used for short-distance communication, such as Infrared Data Association (IrDA) or Bluetooth. The techniques described in this disclosure can also be used in other networks mentioned above.

Figure 2 is a schematic diagram showing a network function according to an embodiment of the present disclosure.

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. A neural network can generally consist of a set of interconnected computational units, which can be referred to as nodes. These nodes may also be referred to as neurons. A neural network consists of at least one node. Nodes (or neurons) that make up neural networks may be interconnected by one or more links.

Within a neural network, one or more nodes connected through a link may form a relative input node and output node relationship. The concepts of input node and output node are relative, and any node in an output node relationship with one node may be in an input node relationship with another node, and vice versa. As described above, input node to output node relationships can be created around links. One or more output nodes can be connected to one input node through a link, and vice versa.

In a relationship between an input node and an output node connected through one link, the value of the data of the output node may be determined based on the data input to the input node. Here, the link connecting the input node and the output node may have a weight. Weights may be variable and may be varied by the user or algorithm in order for the neural network to perform the desired function. For example, when one or more input nodes are connected to one output node by respective links, the output node is set to the values input to the input nodes connected to the output node and the links corresponding to each input node. The output node value can be determined based on the weight.

As described above, in a neural network, one or more nodes are interconnected through one or more links to form an input node and output node relationship within the neural network. The characteristics of the neural network may be determined according to the number of nodes and links within the neural network, the correlation between the nodes and links, and the value of the weight assigned to each link. For example, if the same number of nodes and links exist and two neural networks with different weight values of the links exist, the two neural networks may be recognized as different from each other.

A neural network may consist of a set of one or more nodes. A subset of nodes that make up a neural network can form a layer. Some of the nodes constituting the neural network may form one layer based on the distances from the first input node. For example, a set of nodes with a distance n from the initial input node may constitute n layers. The distance from the initial input node can be defined by the minimum number of links that must be passed to reach the node from the initial input node. However, this definition of a layer is arbitrary for explanation purposes, and the order of a layer within a neural network may be defined in a different way than described above. For example, a layer of nodes may be defined by distance from the final output node.

The initial input node may refer to one or more nodes in the neural network through which data is directly input without going through links in relationships with other nodes. Alternatively, in the relationship between nodes based on links within a neural network, it may refer to nodes that do not have other input nodes connected by links. Similarly, the final output node may refer to one or more nodes that do not have an output node in their relationship with other nodes among the nodes in the neural network. Additionally, hidden nodes may refer to nodes constituting a neural network other than the first input node and the last output node.

The neural network according to an embodiment of the present disclosure is a neural network in which the number of nodes in the input layer may be the same as the number of nodes in the output layer, and the number of nodes decreases and then increases again as it progresses from the input layer to the hidden layer. You can. In addition, the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be less than the number of nodes in the output layer, and the number of nodes decreases as it progresses from the input layer to the hidden layer. there is. In addition, the neural network according to another embodiment of the present disclosure may be a neural network in which the number of nodes in the input layer may be greater than the number of nodes in the output layer, and the number of nodes increases as it progresses from the input layer to the hidden layer. You can. A neural network according to another embodiment of the present disclosure may be a neural network that is a combination of the above-described neural networks.

A deep neural network (DNN) may refer to a neural network that includes multiple hidden layers in addition to the input layer and output layer. Deep neural networks allow you to identify latent structures in data. In other words, it is possible to identify the potential structure of a photo, text, video, voice, or music (e.g., what object is in the photo, what the content and emotion of the text are, what the content and emotion of the voice are, etc.) . Deep neural networks include convolutional neural networks (CNN), recurrent neural networks (RNN), auto encoders, generative adversarial networks (GAN), and restricted Boltzmann machines (RBM). machine), deep belief network (DBN), Q network, U network, Siamese network, Generative Adversarial Network (GAN), etc. The description of the deep neural network described above is only an example and the present disclosure is not limited thereto.

In one embodiment of the present disclosure, the network function may include an autoencoder. An autoencoder may be a type of artificial neural network to output output data similar to input data. The autoencoder may include at least one hidden layer, and an odd number of hidden layers may be placed between input and output layers. The number of nodes in each layer may be reduced from the number of nodes in the input layer to an intermediate layer called the bottleneck layer (encoding), and then expanded symmetrically and reduced from the bottleneck layer to the output layer (symmetrical to the input layer). Autoencoders can perform nonlinear dimensionality reduction. The number of input layers and output layers can be corresponded to the dimension after preprocessing of the input data. In an auto-encoder structure, the number of nodes in the hidden layer included in the encoder may have a structure that decreases as the distance from the input layer increases. If the number of nodes in the bottleneck layer (the layer with the fewest nodes located between the encoder and decoder) is too small, not enough information may be conveyed, so if it is higher than a certain number (e.g., more than half of the input layers, etc.) ) may be maintained.

A neural network may be trained in at least one of supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. Learning of a neural network may be a process of applying knowledge for the neural network to perform a specific operation to the neural network.

Neural networks can be trained to minimize output errors. In neural network learning, learning data is repeatedly input into the neural network, the output of the neural network and the error of the target for the learning data are calculated, and the error of the neural network is transferred from the output layer of the neural network to the input layer in the direction of reducing the error. This is the process of updating the weight of each node in the neural network through backpropagation. In the case of teacher learning, learning data in which the correct answer is labeled in each learning data is used (i.e., labeled learning data), and in the case of non-teacher learning, the correct answer may not be labeled in each learning data. That is, for example, in the case of teacher learning regarding data classification, the learning data may be data in which each learning data is labeled with a category. Labeled training data is input to the neural network, and the error can be calculated by comparing the output (category) of the neural network with the label of the training data. As another example, in the case of non-teachable learning for data classification, the error can be calculated by comparing the input training data with the neural network output. The calculated error is back-propagated in the neural network in the reverse direction (i.e., from the output layer to the input layer), and the connection weight of each node in each layer of the neural network can be updated according to back-propagation. The amount of change in the connection weight of each updated node may be determined according to the learning rate. The neural network's calculation of input data and backpropagation of errors can constitute a learning cycle (epoch). The learning rate may be applied differently depending on the number of repetitions of the learning cycle of the neural network. For example, in the early stages of neural network training, a high learning rate can be used to increase efficiency by allowing the neural network to quickly achieve a certain level of performance, and in the later stages of training, a low learning rate can be used to increase accuracy.

In the learning of neural networks, the training data can generally be a subset of real data (i.e., the data to be processed using the learned neural network), and thus the error for the training data is reduced, but the error for the real data is reduced. There may be an incremental learning cycle. Overfitting is a phenomenon in which errors in actual data increase due to excessive learning on training data. For example, a phenomenon in which a neural network that learned a cat by showing a yellow cat fails to recognize that it is a cat when it sees a non-yellow cat may be a type of overfitting. Overfitting can cause errors in machine learning algorithms to increase. To prevent such overfitting, various optimization methods can be used. To prevent overfitting, methods such as increasing the learning data, regularization, dropout to disable some of the network nodes during the learning process, and use of a batch normalization layer can be applied. You can.

Figure 3 is a flowchart showing a method for learning a neural network model based on the type of data according to an embodiment of the present disclosure.

The computing device 100 according to an embodiment of the present disclosure may directly acquire “data for training a neural network model based on the type of data” or receive it from an external system. The external system may be a server, database, etc. that stores and manages data for learning a neural network model based on the type of data. The computing device 100 may use data directly acquired or received from an external system as “input data for learning a neural network model based on the type of data.”

The computing device 100 may receive original data and determine the type of the original data (S110). For example, the type of original data may include image data or audio data. However, the original data is not limited to examples of image data or audio data, and may include various types of data.

The computing device 100 may determine the ratio of noise based on the type of original data determined in step S110 (S120).

According to an embodiment of the present disclosure, the computing device 100 may determine the ratio of the noise based on the size of data that can be expressed for the type of original data determined in step S110. Specifically, the computing device 100 may determine the ratio of noise so that the size of noise added to the original data is smaller than the minimum size of data that can be expressed for the determined type of original data. For example, if the type of original data is determined to be image data, the original data may have a depth of 8 bits per channel. Accordingly, the computing device 100 may determine the noise ratio so that the size of the noise added to the original data is smaller than 1/(2^8), which is the minimum size of data that can be expressed for image data. A detailed description of this will be provided later with reference to FIGS. 4 and 5.

The computing device 100 may obtain first data by adding noise to the original data based on the noise ratio determined in step S120 (S130). For example, the computing device 100 may acquire first data by adding first noise to the original data based on the determined ratio of noise, and add first data to the original data based on the determined ratio of noise. Second data can be obtained by adding second noise. At this time, the second data including the second noise may include isotropic Gaussian distributed noise, and the computing device 100 partially removes the noise from the Gaussian distributed noise by utilizing the neural network model. The removed data can be obtained as the first data. In addition, the computing device 100 may obtain data from which noise has been completely removed by repeating the process of removing noise using the neural network model, and in this case, data from which noise has been completely removed may be included in the original data, Data containing noise immediately before an image from which noise is completely removed may be included in the first data. In this regard, the specific process in which the second noise is removed from the second data by using a neural network model and the first data is acquired will be described later with reference to FIG. 4. Meanwhile, expressions such as first and second described throughout the present disclosure are only meant to distinguish between components and are not limited thereto.

The computing device 100 may predict noise included in the first data obtained in step S130 using a neural network model (S140). Specifically, the computing device 100 may predict first noise included in the first data using a neural network model, and predict second noise included in the second data using a neural network model. At this time, the neural network model may correspond to a model that obtains data from which part or all of the predicted noise is removed by removing part or all of the predicted noise with respect to Gaussian distributed noise, Diffusion model, score based model, etc. may be included, but are not limited to this and various models may be used in the neural network model.

The computing device 100 may calculate a loss function by comparing the noise predicted through step S140 and the noise added to the original data through step S130 (S150). For example, the computing device 100 calculates a first loss function by comparing the predicted first noise and the first noise added to the original data, and adds the predicted second noise to the first data. A second loss function can be calculated by comparing the added second noise. The calculated loss function can be used in the process of learning a neural network model for generating data, and the specific process will be described later with reference to FIG. 4.

The computing device 100 may train the neural network model based on the loss function calculated through step S150 (S160).

For example, the computing device 100 may train the neural network model based on at least one of the first loss function or the second loss function. Meanwhile, the computing device 100 trains a neural network model to generate data by applying a different noise schedule depending on the type of data, thereby minimizing noise included in data generated of the desired type to improve quality. Technical effects that can generate data can be achieved. Meanwhile, the specific process in which the loss function is calculated and the neural network model is learned based on it will be described later with reference to FIG. 4.

Figure 4 is a neural network model to obtain data from which part or all of the predicted noise is removed by removing part or all of the predicted noise for Gaussian distributed noise according to an embodiment of the present disclosure. This is a schematic diagram to explain the learning process.

Referring to FIG. 4, the computing device 100 according to an embodiment of the present disclosure removes part or all of the isotropic Gaussian distributed noise 10-4, thereby reducing the noise to A neural network model can be trained to obtain data (10-1, 10-2, or 10-3) from which some or all of the data has been removed. At this time, the neural network model may include a conditional noise prediction model. In addition, the conditional noise prediction model may include a U-Net structure where the input and output have the same size, and the noise-containing data x(t) and diffusion time step t may be used as input. And, the diffusion noise included in the noise-containing data x(t) can be predicted and output.

The computing device 100 repeats the process of gradually adding random Gaussian noise over T number of type steps (time steps) to the original data x(0)(10-1) that does not contain noise (from 10-2). direction of 10-3), and as a result, a forward process can be performed to obtain isotropic Gaussian distributed noise x(T) (isotropic gaussian distributed noise) (10-4). At this time, in the example of FIG. 4, image data is disclosed as an example, but the type of the original data x(0)(10-1) may include various examples such as audio data in addition to image data. Meanwhile, the forward process according to an embodiment of the present disclosure can be performed using the following equation.

[Equation 1]

In Equation 1 above, can be used as a hyperparameter in the process of calculating the diffusion coefficient, can be set to any value, 0< < < … < It can be set to a value of <1. for example This has a value of 0.0001 can have a value of 0.02, from until The value of may increase linearly or according to a cosine function, and T, which means the total number of diffusion steps, may be set to 1000. but, It is only an example that increases linearly or according to a cosine function, and according to embodiments of the present disclosure, depending on the type of the original data The amount of increase may be determined differently, and detailed explanations are provided below through [Equation 2] and FIG. 5. also, may mean random Gaussian distributed noise. The general formula for expressing the noise-containing data x(t) at time step t with the noise-free original data x(0)(10-1) and the included diffusion noise is as follows: can be expressed.

[Equation 2]

According to an embodiment of the present disclosure, equations (1) and (2) in Equation 2 are the diffusion coefficient It is a way to express the specific meaning of . In equation (1) of Equation 2 above, the diffusion coefficient at a specific time step t The hyperparameter above is 1 It can be calculated by subtracting the diffusion coefficient in equation (2) of Equation 2 above. may mean the diffusion coefficient sequentially accumulated from time step 1 to t. Therefore, equation (3) in Equation 2 is the data x(t) containing the noise at time step t, the original data x(0)(10-1) without the noise, and the diffusion coefficient ( ) and random Gaussian distributed noise ( ) is expressed as a formula for . In addition, equation (4) in Equation 2 is the noise ratio n(t) and signal ratio s(t) determined based on the type of the original data 10-1 according to an embodiment of the present disclosure. It can mean. Therefore, data x(t) containing noise at time step t can be expressed based on the noise ratio n(t) and the signal ratio s(t). Specifically, the data x(t) containing noise at time step t becomes the original data x(0)(10-1) without noise as the ratio of the original data x(0)(10-1) (i.e., the ratio of signals s(t)) increases. It gets closer to the shape of the data, and random Gaussian distribution noise ( ) (i.e., the ratio of noise n(t)) increases, the closer it can get to the form of random Gaussian distribution noise.

For example, the computing device 100 adds random Gaussian noise little by little over T number of type steps (time steps) to the original data x(0)(10-1) that does not contain noise. First data 10-2 is obtained by adding first noise to the original data 10-1 based on the ratio of noise, and the first data 10-2 is obtained based on the determined ratio of noise. Second data 10-3 can be obtained by adding second noise to . At this time, in the example of FIG. 4, image data is disclosed as an example, but the type of the original data x(0)(10-1) may include various examples such as audio data in addition to image data.

Through this, the computing device 100 calculates T based on “the ratio of noise determined through equation (4) in Equation 2 above” with respect to the original data The process of adding random Gaussian noise is repeated over time steps (in the direction from 10-2 to 10-3), resulting in isotropic Gaussian distributed noise x(T) (10-3). 4) A forward process can be performed to obtain. However, the forward process is not limited to the example of Equation 2 above, and various processes that add noise to data may be included in the forward process.

In addition, the computing device 100 generates random Gaussian noise little by little over T type steps (time steps) with respect to the isotropic Gaussian distributed noise x(T) (10-4) in the opposite direction to the forward process. Repeat the process of removing (in the direction from 10-3 to 10-2), and learn the neural network model to perform the reverse process to obtain original data x(0)(10-1) that does not contain noise. You can do it. In this regard, the formula representing the reverse execution process can be expressed as follows.

[Equation 3]

In Equation 3, equation (1) is the noise prediction result predicted by the neural network model for “data containing noise x(t)” ( ) is a formula that represents the reverse process of obtaining “data x(t-1) from the previous stage from which some of the noise has been removed.” For example, the computing device 100 removes the noise prediction result predicted by the neural network model for “second data x(2)(10-3) containing second noise” to determine “the second noise is The first data x(1)(10-2)” of the previous step that was removed can be obtained. In Equation 3 above, equation (2) is the diffusion coefficient at the current time step t means, and in equation (3) in equation 3 above, means the dispersion parameter, diffusion coefficient It can be calculated based on . However, the reverse process is not limited to [Equation 3], and various processes for removing noise from data containing noise may be included in the reverse process.

Specifically, the computing device 100 inputs data x(t) and time step t containing noise into the neural network model, and calculates the noise prediction result predicted by the neural network model and the diffusion noise actually included. By comparing, a loss function can be calculated, and gradient descent can be performed according to the loss function to learn the neural network model. For example, the loss function calculated by comparing the noise prediction result predicted by the neural network model with the diffusion noise actually included can be expressed as the following equation.

[Equation 4]

In [Equation 4] above, the loss function is the diffusion noise actually included ( ) and the noise prediction result predicted by the neural network model ( ) can be calculated by comparing. For example, the computing device 100 predicts “the first noise included in the first data x(1)(10-2)” and predicts the predicted first noise and the original data x(0)( A first loss function may be calculated by comparing the first noise added to 10-1), and the computing device 100 may calculate “second noise included in the second data x(2)(10-3)” and calculate a second loss function by comparing the predicted second noise and the second noise added to the first data x(1)(10-2). However, the loss function is not limited to the example of Equation 4, and may include various loss functions calculated by comparing the noise prediction result and the actually included diffusion noise.

Additionally, the neural network model predicts the diffusion noise included in the data x(2)(10-3) including the second noise, and removes the predicted second noise to produce “data x from which the second noise has been removed.” (1)(10-2)” can be learned to obtain. For example, the computing device 100 may train the neural network model based on at least one of the first loss function or the second loss function. In addition, the neural network model predicts the diffusion noise included in the “data x(t) containing the noise,” repeats the process of removing the predicted diffusion noise one or more times, and repeats the process of “data containing the noise By removing all the diffusion noise included in “(t)”, the original data x(0)(10-1) with all the noise removed can be learned to be obtained. Meanwhile, in the process of learning to remove diffusion noise from data containing noise using the neural network model, the ratio of noise is determined based on the type of original data, so that the neural network model is used to generate better quality data. It can be learned, and the specific process for this is described later with reference to FIG. 5.

Figure 5 is a schematic diagram illustrating a process of determining the type of received original data and determining the noise ratio based on the determined type of original data according to an embodiment of the present disclosure.

Referring to FIG. 5, the computing device 100 may determine the ratio of noise based on the size of data that can be expressed for the determined type of original data 10-1. Specifically, the computing device 100 may determine the ratio of noise so that the size of noise added to the original data is smaller than the minimum size of data that can be expressed for the determined type of original data. For example, if the type of the original data 10-1 is determined to be image data, the original data 10-1 may have a depth of 8 bits per channel. Therefore, the computing device 100 sets the size of the noise added to the original data 10-1 to be smaller than 1/(2^8) (=about 0.00039), which is the minimum size of data that can be expressed for image data. The ratio of noise (20) can be determined.

Meanwhile, according to another embodiment of the present disclosure, when the type of the original data 10-1 is determined to be audio data, the original data may have a depth of 16 bits per sample. there is. Therefore, the computing device 100 sets the size of the noise added to the original data 10-1 to be smaller than 1/(2^16) (=about 0.000015), which is the minimum size of data that can be expressed for audio data. The ratio of noise (30) can be determined. Accordingly, the noise ratio 30 when the original data is audio data may be determined to be smaller than the noise ratio 20 when the original data is image data. For example, the noise ratio can be determined through the following equation.

[Equation 5]

Specifically, according to an embodiment of the present disclosure, referring to equation (1) in [Equation 5], the noise ratio n(t) can be determined to increase exponentially, and in [Equation 5] Referring to equation (2), the signal rate s(t) can be determined to decrease exponentially. At this time, r may mean a hyperparameter that determines the shape of the curve of the noise ratio n(t) and the signal ratio s(t). However, the above [Equation 5] is only an example, and in addition, the noise ratio and signal ratio can be determined in various ways. Through this, the neural network model can be learned so that the results generated through the neural network model have little influence of noise even in data domains (eg, audio data, etc.) that are relatively sensitive to noise compared to image data. Meanwhile, the specific process of acquiring the determined type of data by utilizing a neural network model learned in consideration of the loss function associated with the ratio of noise determined according to the type of data will be described later with reference to FIG. 6.

6 illustrates determining the type of data to be generated according to an embodiment of the present disclosure, and obtaining the determined type of data by utilizing a neural network model learned by considering the loss function associated with the ratio of noise determined according to the type of data. This is a flowchart showing the process.

The computing device 100 according to an embodiment of the present disclosure may determine the type of data to be generated (S210). For example, the computing device 100 may determine the type of data to be generated as audio data. However, audio data is only an example and various types of data can be determined as the type of data to be generated.

In addition, the computing device 100 uses a neural network model to remove part or all of the noise for Gaussian distributed noise, and provides data of the type determined through step S210 from which part or all of the noise has been removed. Can be obtained (S220). At this time, the neural network model may correspond to a model learned by considering the loss function associated with the ratio of noise determined according to the type of data. In this regard, the above-described embodiments of the present disclosure may be referred to for the specific process in which the neural network model is learned by considering the “loss function associated with the ratio of noise determined according to the type of data.” For example, the computing device 100 determines the type of data to be generated as audio, uses the neural network model to remove noise for Gaussian distributed noise, and obtains audio data including noise of 0.00001. can do. At this time, since the size of the noise included in the audio data acquired using the neural network model (i.e., 0.00001) is smaller than the minimum size of data that can be expressed for audio (about 0.000015), the generated result has little influence of noise. There may not be. Therefore, the computing device 100 generates data using the neural network model learned by considering the loss function associated with the ratio of noise determined according to the type of data, thereby generating better quality data with less influence of noise. Possible technical effects can be achieved.

According to an embodiment of the present disclosure, a computer-readable medium storing a data structure is disclosed. Data structure can refer to the organization, management, and storage of data to enable efficient access and modification of data. Data structure can refer to the organization of data to solve a specific problem (e.g., retrieving data, storing data, or modifying data in the shortest possible time). A data structure may be defined as a physical or logical relationship between data elements designed to support a specific data processing function. Logical relationships between data elements may include connection relationships between user-defined data elements. Physical relationships between data elements may include actual relationships between data elements that are physically stored in a computer-readable storage medium (e.g., a persistent storage device). A data structure may specifically include a set of data, relationships between data, and functions or instructions applicable to the data. Effectively designed data structures allow computing devices to perform computations while minimizing the use of the computing device's resources. Specifically, computing devices can increase the efficiency of operations, reading, insertion, deletion, comparison, exchange, and search through effectively designed data structures.

Data structures can be divided into linear data structures and non-linear data structures depending on the type of data structure. A linear data structure may be a structure in which only one piece of data is connected to another piece of data. Linear data structures may include List, Stack, Queue, and Deque. A list can refer to a set of data that has an internal order. The list may include a linked list. A linked list may be a data structure in which data is connected in such a way that each data is connected in a single line with a pointer. In a linked list, a pointer may contain connection information to the next or previous data. Depending on its form, a linked list can be expressed as a singly linked list, a doubly linked list, or a circularly linked list. A stack may be a data listing structure that allows limited access to data. A stack can be a linear data structure in which data can be processed (for example, inserted or deleted) at only one end of the data structure. Data stored in the stack may have a data structure (LIFO-Last in First Out) where the later it enters, the sooner it comes out. A queue is a data listing structure that allows limited access to data. Unlike the stack, it can be a data structure (FIFO-First in First Out) where data stored later is released later. A deck can be a data structure that can process data at both ends of the data structure.

A non-linear data structure may be a structure in which multiple pieces of data are connected behind one piece of data. Nonlinear data structures may include graph data structures. A graph data structure can be defined by vertices and edges, and an edge can include a line connecting two different vertices. Graph data structure may include a tree data structure. A tree data structure may be a data structure in which there is only one path connecting two different vertices among a plurality of vertices included in the tree. In other words, it may be a data structure that does not form a loop in the graph data structure.

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. Below, it is described in a unified manner as a neural network. Data structures may include neural networks. And the data structure including the neural network may be stored in a computer-readable medium. Data structures including neural networks also include data preprocessed for processing by a neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data acquired from the neural network, activation functions associated with each node or layer of the neural network, neural network It may include a loss function for learning. A data structure containing a neural network may include any of the components disclosed above. In other words, the data structure including the neural network includes preprocessed data for processing by the neural network, data input to the neural network, weights of the neural network, hyperparameters of the neural network, data acquired from the neural network, activation functions associated with each node or layer of the neural network, neural network It may be configured to include all or any combination of the loss function for learning. In addition to the configurations described above, a data structure containing a neural network may include any other information that determines the characteristics of the neural network. Additionally, the data structure may include all types of data used or generated in the computational process of a neural network and is not limited to the above. Computer-readable media may include computer-readable recording media and/or computer-readable transmission media. A neural network can generally consist of a set of interconnected computational units, which can be referred to as nodes. These nodes may also be referred to as neurons. A neural network consists of at least one node.

The data structure may include data input to the neural network. A data structure containing data input to a neural network may be stored in a computer-readable medium. Data input to the neural network may include learning data input during the neural network learning process and/or input data input to the neural network on which training has been completed. Data input to the neural network may include data that has undergone pre-processing and/or data subject to pre-processing. Preprocessing may include a data processing process to input data into a neural network. Therefore, the data structure may include data subject to preprocessing and data generated by preprocessing. The above-described data structure is only an example and the present disclosure is not limited thereto.

The data structure may include the weights of the neural network. (In this specification, weights and parameters may be used with the same meaning.) And the data structure including the weights of the neural network may be stored in a computer-readable medium. A neural network may include multiple weights. Weights may be variable and may be varied by the user or algorithm in order for the neural network to perform the desired function. For example, when one or more input nodes are connected to one output node by respective links, the output node is set to the values input to the input nodes connected to the output node and the links corresponding to each input node. Based on the weight, the data value output from the output node can be determined. The above-described data structure is only an example and the present disclosure is not limited thereto.

As an example and not a limitation, the weights may include weights that are changed during the neural network learning process and/or weights for which neural network learning has been completed. Weights that change during the neural network learning process may include weights that change at the start of the learning cycle and/or weights that change during the learning cycle. Weights for which neural network training has been completed may include weights for which a learning cycle has been completed. Therefore, the data structure including the weights of the neural network may include weights that are changed during the neural network learning process and/or the data structure including the weights for which neural network learning has been completed. Therefore, the above-mentioned weights and/or combinations of each weight are included in the data structure including the weights of the neural network. The above-described data structure is only an example and the present disclosure is not limited thereto.

The data structure including the weights of the neural network may be stored in a computer-readable storage medium (e.g., memory, hard disk) after going through a serialization process. Serialization can be the process of converting a data structure into a form that can be stored on the same or a different computing device and later reorganized and used. Computing devices can transmit and receive data over a network by serializing data structures. Data structures containing the weights of a serialized neural network can be reconstructed on the same computing device or on a different computing device through deserialization. The data structure including the weights of the neural network is not limited to serialization. Furthermore, the data structure including the weights of the neural network is a data structure to increase computational efficiency while minimizing the use of computing device resources (e.g., in non-linear data structures, B-Tree, Trie, m-way search tree, AVL tree, Red-Black Tree) may be included. The foregoing is merely an example and the present disclosure is not limited thereto.

The data structure may include hyper-parameters of a neural network. And the data structure including the hyperparameters of the neural network can be stored in a computer-readable medium. A hyperparameter may be a variable that can be changed by the user. Hyperparameters include, for example, learning rate, cost function, number of learning cycle repetitions, weight initialization (e.g., setting the range of weight values subject to weight initialization), Hidden Unit. It may include a number (e.g., number of hidden layers, number of nodes in hidden layers). The above-described data structure is only an example and the present disclosure is not limited thereto.

7 is a brief, general schematic diagram of an example computing environment in which embodiments of the present disclosure may be implemented.

Although the disclosure has generally been described above as being capable of being implemented by a computing device, those skilled in the art will understand that the disclosure can be implemented in combination with computer-executable instructions and/or other program modules that can be executed on one or more computers and/or hardware. It will be well known that it can be implemented as a combination of software.

Typically, program modules include routines, programs, components, data structures, etc. that perform specific tasks or implement specific abstract data types. Additionally, those skilled in the art will understand that the methods of the present disclosure can be used on single-processor or multiprocessor computer systems, minicomputers, mainframe computers, as well as personal computers, handheld computing devices, microprocessor-based or programmable consumer electronics, etc. It will be appreciated that other computer system configurations may be implemented, including, but not limited to, each of which may operate in connection with one or more associated devices.

The described embodiments of the disclosure can also be practiced in distributed computing environments where certain tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

Computers typically include a variety of computer-readable media. Computer-readable media can be any medium that can be accessed by a computer, and such computer-readable media includes volatile and non-volatile media, transitory and non-transitory media, removable and non-transitory media. Includes removable media. By way of example, and not limitation, computer-readable media may include computer-readable storage media and computer-readable transmission media. Computer-readable storage media refers to volatile and non-volatile media, transient and non-transitory media, removable and non-removable, implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Includes media. Computer readable storage media may include RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital video disk (DVD) or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage. This includes, but is not limited to, a device, or any other medium that can be accessed by a computer and used to store desired information.

A computer-readable transmission medium typically implements computer-readable instructions, data structures, program modules, or other data on a modulated data signal, such as a carrier wave or other transport mechanism. Includes all information delivery media. The term modulated data signal refers to a signal in which one or more of the characteristics of the signal have been set or changed to encode information within the signal. By way of example, and not limitation, computer-readable transmission media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared, and other wireless media. Combinations of any of the above are also intended to be included within the scope of computer-readable transmission media.

An example environment 1100 is shown that implements various aspects of the present disclosure, including a computer 1102, which includes a processing unit 1104, a system memory 1106, and a system bus 1108. do. System bus 1108 couples system components, including but not limited to system memory 1106, to processing unit 1104. Processing unit 1104 may be any of a variety of commercially available processors. Dual processors and other multiprocessor architectures may also be used as processing unit 1104.

System bus 1108 may be any of several types of bus structures that may further be interconnected to a memory bus, peripheral bus, and local bus using any of a variety of commercial bus architectures. System memory 1106 includes read only memory (ROM) 1110 and random access memory (RAM) 1112. The basic input/output system (BIOS) is stored in non-volatile memory 1110, such as ROM, EPROM, and EEPROM, and is a basic input/output system that helps transfer information between components within the computer 1102, such as during startup. Contains routines. RAM 1112 may also include high-speed RAM, such as static RAM, for caching data.

Computer 1102 may also include an internal hard disk drive (HDD) 1114 (e.g., EIDE, SATA)—the internal hard disk drive 1114 may also be configured for external use within a suitable chassis (not shown). Yes - a magnetic floppy disk drive (FDD) 1116 (e.g., for reading from or writing to a removable diskette 1118), and an optical disk drive 1120 (e.g., a CD-ROM for reading the disk 1122 or reading from or writing to other high-capacity optical media such as DVDs). Hard disk drive 1114, magnetic disk drive 1116, and optical disk drive 1120 are connected to system bus 1108 by hard disk drive interface 1124, magnetic disk drive interface 1126, and optical drive interface 1128, respectively. ) can be connected to. The interface 1124 for implementing an external drive includes at least one or both of Universal Serial Bus (USB) and IEEE 1394 interface technologies.

These drives and their associated computer-readable media provide non-volatile storage of data, data structures, computer-executable instructions, and the like. For computer 1102, drive and media correspond to storing any data in a suitable digital format. Although the description of computer-readable media above refers to removable optical media such as HDDs, removable magnetic disks, and CDs or DVDs, those of ordinary skill in the art would also recognize zip drives, magnetic cassettes, flash memory cards, and cartridges. It will be appreciated that other types of computer-readable media may also be used in the exemplary operating environment, and that any such media may contain computer-executable instructions for performing the methods of the present disclosure. .

A number of program modules may be stored in the drive and RAM 1112, including an operating system 1130, one or more application programs 1132, other program modules 1134, and program data 1136. All or portions of the operating system, applications, modules and/or data may also be cached in RAM 1112. It will be appreciated that the present disclosure may be implemented on various commercially available operating systems or combinations of operating systems.

A user may enter commands and information into computer 1102 through one or more wired/wireless input devices, such as a keyboard 1138 and a pointing device such as mouse 1140. Other input devices (not shown) may include microphones, IR remote controls, joysticks, game pads, stylus pens, touch screens, etc. These and other input devices are connected to the processing unit 1104 through an input device interface 1142, which is often connected to the system bus 1108, but may also include a parallel port, an IEEE 1394 serial port, a game port, a USB port, an IR interface, It can be connected by other interfaces, etc.

A monitor 1144 or other type of display device is also connected to system bus 1108 through an interface, such as a video adapter 1146. In addition to monitor 1144, computers typically include other peripheral output devices (not shown) such as speakers, printers, etc.

Computer 1102 may operate in a networked environment using logical connections to one or more remote computers, such as remote computer(s) 1148, via wired and/or wireless communications. Remote computer(s) 1148 may be a workstation, computing device computer, router, personal computer, portable computer, microprocessor-based entertainment device, peer device, or other conventional network node, and is generally connected to computer 1102. For simplicity, only memory storage device 1150 is shown, although it includes many or all of the components described. The logical connections depicted include wired/wireless connections to a local area network (LAN) 1152 and/or a larger network, such as a wide area network (WAN) 1154. These LAN and WAN networking environments are common in offices and companies and facilitate enterprise-wide computer networks, such as intranets, all of which can be connected to a worldwide computer network, such as the Internet.

When used in a LAN networking environment, computer 1102 is connected to local network 1152 through wired and/or wireless communication network interfaces or adapters 1156. Adapter 1156 may facilitate wired or wireless communication to LAN 1152, which also includes a wireless access point installed thereon for communicating with wireless adapter 1156. When used in a WAN networking environment, the computer 1102 may include a modem 1158 or be connected to a communicating computing device on the WAN 1154 or to establish communications over the WAN 1154, such as via the Internet. Have other means. Modem 1158, which may be internal or external and a wired or wireless device, is coupled to system bus 1108 via serial port interface 1142. In a networked environment, program modules described for computer 1102, or portions thereof, may be stored in remote memory/storage device 1150. It will be appreciated that the network connections shown are exemplary and that other means of establishing a communications link between computers may be used.

Computer 1102 may be associated with any wireless device or object deployed and operating in wireless communications, such as a printer, scanner, desktop and/or portable computer, portable data assistant (PDA), communications satellite, wirelessly detectable tag. Performs actions to communicate with any device or location and telephone. This includes at least Wi-Fi and Bluetooth wireless technologies. Accordingly, communication may be a predefined structure as in a conventional network or may simply be ad hoc communication between at least two devices.

Wi-Fi (Wireless Fidelity) allows connection to the Internet, etc. without wires. Wi-Fi is a wireless technology, like cell phones, that allows these devices, such as computers, to send and receive data indoors and outdoors, anywhere within the coverage area of a cell tower. Wi-Fi networks use wireless technology called IEEE 802.11 (a, b, g, etc.) to provide secure, reliable, and high-speed wireless connections. Wi-Fi can be used to connect computers to each other, to the Internet, and to wired networks (using IEEE 802.3 or Ethernet). Wi-Fi networks can operate in the unlicensed 2.4 and 5 GHz wireless bands, for example, at data rates of 11 Mbps (802.11a) or 54 Mbps (802.11b), or in products that include both bands (dual band). .

Those skilled in the art will understand that information and signals may be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols and chips that may be referenced in the above description include voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields. It can be expressed by particles or particles, or any combination thereof.

Those skilled in the art will understand that the various illustrative logical blocks, modules, processors, means, circuits and algorithm steps described in connection with the embodiments disclosed herein may be used in electronic hardware, (for convenience) It will be understood that it may be implemented by various forms of program or design code (referred to herein as software) or a combination of both. To clearly illustrate this interoperability of hardware and software, various illustrative components, blocks, modules, circuits and steps have been described above generally with respect to their functionality. Whether this functionality is implemented as hardware or software depends on the specific application and design constraints imposed on the overall system. A person skilled in the art of this disclosure may implement the described functionality in various ways for each specific application, but such implementation decisions should not be construed as departing from the scope of this disclosure.

The various embodiments presented herein may be implemented as a method, apparatus, or article of manufacture using standard programming and/or engineering techniques. The term article of manufacture includes a computer program, carrier, or media accessible from any computer-readable storage device. For example, computer-readable storage media include magnetic storage devices (e.g., hard disks, floppy disks, magnetic strips, etc.), optical disks (e.g., CDs, DVDs, etc.), smart cards, and flash. Includes, but is not limited to, memory devices (e.g., EEPROM, cards, sticks, key drives, etc.). Additionally, various storage media presented herein include one or more devices and/or other machine-readable media for storing information.

It is to be understood that the specific order or hierarchy of steps in the processes presented is an example of illustrative approaches. It is to be understood that the specific order or hierarchy of steps in processes may be rearranged within the scope of the present disclosure, based on design priorities. The appended method claims present elements of the various steps in a sample order but are not meant to be limited to the particular order or hierarchy presented.

The description of the presented embodiments is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments without departing from the scope of the disclosure. Thus, the present disclosure is not limited to the embodiments presented herein but is to be interpreted in the broadest scope consistent with the principles and novel features presented herein.

Claims (19)

1. A method for training a neural network model to generate data, performed by one or more processors of a computing device, comprising:
Receiving original data and determining a type of the original data;
determining a noise ratio so that the size of noise added to the original data is smaller than the minimum size of data that can be expressed for the determined type of original data;
acquiring first data by adding noise to the original data based on the determined noise ratio;
predicting noise included in the first data using a neural network model;
Comparing the predicted noise and noise added to the original data to calculate a loss function; and
training the neural network model based on the loss function;
Including,
method.
According to claim 1,
The neural network model is,
A model for obtaining data from which part or all of the predicted noise has been removed by removing part or all of the predicted noise for Gaussian distributed noise,
method.
delete delete According to claim 1,
The types of original data are:
Containing image data or audio data,
method.
According to claim 1,
Obtaining first data by adding noise to the original data based on the determined noise ratio,
acquiring first data by adding first noise to the original data based on the determined noise ratio; and
acquiring second data by adding second noise to the first data based on the determined noise ratio;
Including,
method.
According to claim 6,
The step of predicting noise included in the first data using the neural network model includes:
predicting first noise included in the first data using the neural network model; and
Predicting second noise included in the second data using the neural network model,
The step of calculating a loss function by comparing the predicted noise and the noise added to the original data includes:
Comparing the predicted first noise and the first noise added to the original data to calculate a first loss function; and
Comparing the predicted second noise and the second noise added to the first data to calculate a second loss function,
The step of learning the neural network model based on the loss function is,
Comprising training the neural network model based on at least one of the first loss function or the second loss function,
method.
A computer program stored in a computer-readable storage medium, wherein the computer program, when executed by one or more processors, causes the one or more processors to perform operations for training a neural network model for generating data, the operations being: :
Receiving original data and determining a type of the original data;
determining a noise ratio so that the size of noise added to the original data is smaller than the minimum size of data that can be expressed for the determined type of original data;
acquiring first data by adding noise to the original data based on the determined noise ratio;
Predicting noise included in the first data using a neural network model;
calculating a loss function by comparing the predicted noise and noise added to the original data; and
An operation of training the neural network model based on the loss function;
Including,
A computer program stored on a computer-readable storage medium.
delete delete According to claim 8,
The operation of acquiring first data by adding noise to the original data based on the determined noise ratio,
acquiring first data by adding first noise to the original data based on the determined noise ratio; and
acquiring second data by adding second noise to the first data based on the determined noise ratio;
Including,
A computer program stored on a computer-readable storage medium.
According to claim 11,
The operation of predicting noise included in the first data using the neural network model includes:
Predicting first noise included in the first data using the neural network model; and
An operation of predicting second noise included in the second data using the neural network model,
The operation of calculating a loss function by comparing the predicted noise and the noise added to the original data includes:
calculating a first loss function by comparing the predicted first noise and the first noise added to the original data; and
Comprising the operation of calculating a second loss function by comparing the predicted second noise and the second noise added to the first data,
The operation of training the neural network model based on the loss function is:
Comprising the operation of training the neural network model based on at least one of the first loss function or the second loss function,
A computer program stored on a computer-readable storage medium.
As a computing device,
at least one processor; and
Memory
Including,
The at least one processor,
Receiving original data and determining the type of the original data;
determining a noise ratio so that the size of noise added to the original data is smaller than the minimum size of data that can be expressed for the determined type of original data;
Obtaining first data by adding noise to the original data based on the determined noise ratio;
predict noise included in the first data using a neural network model;
Compare the predicted noise and the noise added to the original data to calculate a loss function; and
configured to train the neural network model based on the loss function,
Computing device.
delete delete According to claim 13,
The process of acquiring first data by adding noise to the original data based on the determined noise ratio includes:
Obtaining first data by adding first noise to the original data based on the determined noise ratio; and
Comprising the process of acquiring second data by adding second noise to the first data based on the determined ratio of noise,
Computing device.
According to claim 16,
The process of predicting noise included in the first data using the neural network model is:
predicting first noise included in the first data using the neural network model; and
A process of predicting second noise included in the second data using the neural network model,
The process of calculating a loss function by comparing the predicted noise and the noise added to the original data is,
Compare the predicted first noise and the first noise added to the original data to calculate a first loss function; and
Comprising the process of calculating a second loss function by comparing the predicted second noise and the second noise added to the first data,
The process of learning the neural network model based on the loss function is,
Including the process of training the neural network model based on at least one of the first loss function or the second loss function,
Computing device.
A computer-readable storage medium in which the parameters of a neural network, at least some of which are updated during the learning process, are recorded,
The neural network performs the following steps based at least in part on the parameters,
The above steps are:
Receiving original data and determining a type of the original data;
determining a noise ratio so that the size of noise added to the original data is smaller than the minimum size of data that can be expressed for the determined type of original data;
acquiring first data by adding noise to the original data based on the determined noise ratio;
predicting noise included in the first data using a neural network model;
Comparing the predicted noise and noise added to the original data to calculate a loss function; and
training the neural network model based on the loss function;
Including,
Computer readable storage medium.
A method of generating data using a neural network model performed by a computing device, comprising:
determining the type of data to be generated; and
Using a neural network model based on a diffusion model, removing part or all of the noise for Gaussian distributed noise, and obtaining the determined type of data from which part or all of the noise has been removed. ;
Including,
The neural network model based on the diffusion model is,
It is a model learned considering the loss function associated with the ratio of noise determined according to the type of data,
The determined noise ratio is,
The size of the noise added to the original data during the learning process is determined to be smaller than the minimum size of data that can be expressed for the determined type of data.
method.