CN114781628A - Memristor noise-based data enhancement method and device, electronic equipment and medium - Google Patents
Memristor noise-based data enhancement method and device, electronic equipment and medium Download PDFInfo
- Publication number
- CN114781628A CN114781628A CN202210325808.1A CN202210325808A CN114781628A CN 114781628 A CN114781628 A CN 114781628A CN 202210325808 A CN202210325808 A CN 202210325808A CN 114781628 A CN114781628 A CN 114781628A
- Authority
- CN
- China
- Prior art keywords
- data
- memristor
- mapping
- noise
- training
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000013507 mapping Methods 0.000 claims abstract description 69
- 238000012549 training Methods 0.000 claims description 61
- 230000006870 function Effects 0.000 claims description 20
- 238000013528 artificial neural network Methods 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 7
- 230000002708 enhancing effect Effects 0.000 claims description 6
- 238000005516 engineering process Methods 0.000 abstract description 6
- 230000008569 process Effects 0.000 description 13
- 238000010586 diagram Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 7
- 230000008901 benefit Effects 0.000 description 6
- 230000007547 defect Effects 0.000 description 6
- 230000009466 transformation Effects 0.000 description 5
- 230000000875 corresponding effect Effects 0.000 description 4
- 230000001965 increasing effect Effects 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 210000004556 brain Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000003631 expected effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
- G06N3/063—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using electronic means
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Neurology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Noise Elimination (AREA)
Abstract
The application discloses a memristor noise-based data enhancement method, a memristor noise-based data enhancement device, electronic equipment and a medium, wherein the method comprises the following steps: determining a mapping relation representing a relation between input data and output data; mapping a mapping network corresponding to the mapping relation to the target memristor array based on the mapping relation; and inputting the input data to the mapped target memristor array, and applying random noise to the target memristor array to obtain data-enhanced output data. According to the embodiment of the application, random noise of the memristor is utilized to enhance data, the enhanced data have diversity and randomness, and the problems that in the related technology, a data set applicable to an offline data enhancement mode is small, an online data enhancement mode is long in time consumption and low in efficiency, and the data enhancement mode is single are solved.
Description
Technical Field
The present disclosure relates to the field of data enhancement technologies, and in particular, to a method and an apparatus for enhancing data based on memristor noise, an electronic device, and a storage medium.
Background
Artificial intelligence is a technological science for researching and developing theories, methods, techniques and application systems for simulating, extending and expanding human behaviors, is a subject for researching and enabling a computer to simulate certain thinking processes and intelligent behaviors (such as learning, reasoning, thinking, planning and the like) of a human, and mainly comprises the principle that the computer realizes intelligence and the computer similar to human brain intelligence is manufactured, so that the computer can realize higher-level application. Generally, to obtain an ideal artificial intelligence model, a large amount of labeled data is used to make a model undergo supervised learning. The model can learn a function, namely a model parameter, from a given training data set, and when new data comes, a result can be predicted according to the function, so that the purpose of prediction is achieved.
The predictive effect of supervised learning is largely positively correlated with the amount and diversity of data accepted during the training phase. With the increasing scale of neural networks, the demand for data volume and data diversity is also increasing. However, compared with the huge data volume requirement, the existing data set cannot meet the requirement. Therefore, one solution to this problem is to use data enhancement, i.e., to generate new training data by performing certain transformations on the limited data, so as to achieve the purpose of data expansion. In addition to expanding data volume and data diversity, data enhancement can be used to solve the problem of class imbalance in classification tasks, such as adjusting the ratio of positive and negative samples by data enhancement.
There are many implementation methods for data enhancement, and the following takes image data as an example, and the data enhancement is implemented by performing regular transformation such as cutting, turning, rotating and the like on the existing data. Data enhancement is generally divided into offline data enhancement and online data enhancement. The offline data enhancement means that the expanded data set is cached after the data set is processed for training and reasoning of the model. The online data enhancement means that only the data of the current training batch is changed in the model training and reasoning process. One of the most similar on-line data enhancement schemes is referred to as an autoencoder, which first compression encodes the data into a specific vector and then adds a sample of gaussian noise to the specific vector. The data enhancement scheme that restores the original image to the decoder is called the auto-encoder. The method has the core that after data passes through a neural network, random Gaussian disturbance is added, so that the data is transformed. Since random disturbance is added, the robustness of the training model can be improved.
In which, the offline data enhancement directly processes the data set, and the number of data becomes the enhancement factor multiplied by the number of the original data set. The method has the advantages that model training and reasoning time length does not need to be increased, and the preprocessed data can be used for different tasks of a plurality of models. However, the off-line data enhancement scheme has obvious disadvantages, and has high requirements on chip cache because the data set is directly processed. This scheme is typically used for smaller data sets.
On-line data enhancement is to transform only the data of the current use batch in the model training process. During the training process, the data that has been used is not saved. The scheme has the advantages that no extra buffer is needed, and data can be deleted immediately in the model training and reasoning process. However, this solution also has a significant disadvantage, and when a more complex transformation is encountered, the online data enhancement will lengthen the duration of model training and reasoning. Taking the self-encoder scheme mentioned above as an example, the process of generating random numbers takes a lot of time. Randomly generating a large number of random numbers for each batch would severely increase the time cost of model training and reasoning.
Transforming an image using a single transformation means, such as translation, rotation, scaling, etc., is one of the commonly used schemes for image data enhancement. The main defect is that the generated images lack diversity and randomness, so that the improvement on the robustness of the model cannot achieve the expected effect.
Disclosure of Invention
The application provides a memristor noise-based data enhancement method and device, electronic equipment and a storage medium, random noise of a memristor is utilized to enhance data, the enhanced data have diversity and randomness, and the problems that in the related art, an offline data enhancement mode is applicable to a small data set, an online data enhancement mode is long in time consumption, low in efficiency and single in data enhancement mode are solved.
The embodiment of the first aspect of the application provides a data enhancement method based on memristor noise, which comprises the following steps: determining a mapping relation representing a relation between input data and output data; mapping a mapping network corresponding to the mapping relation to a target memristor array based on the mapping relation; and inputting the input data to the mapped target memristor array, and applying random noise to the target memristor array to obtain the output data after data enhancement.
Optionally, in an embodiment of the present application, the method further includes: and performing neural network training by using the mapping relation as training data, and stopping training when a training termination condition is met to obtain the mapping network.
Optionally, in an embodiment of the present application, the inputting the input data to the mapped target memristor array, and obtaining the data-enhanced output data after applying random noise to the target memristor array includes: inputting a voltage signal of the input data to the mapped target memristor array; outputting a current signal of the input data after applying random noise through the memristor array; and converting the current signal to obtain the output data after the data enhancement.
Optionally, in an embodiment of the present application, the training termination condition includes: the loss function of the mapping network is smaller than a preset threshold value; and/or the error between the input data and the output data is smaller than a preset error threshold value.
The embodiment of the second aspect of the present application provides a data enhancement device based on memristor noise, including: the acquisition module is used for determining a mapping relation representing the relation between the input data and the output data; the mapping module is used for mapping a mapping network corresponding to the mapping relation to a target memristor array based on the mapping relation; and the enhancing module is used for inputting the input data to the mapped target memristor array and applying random noise to the target memristor array to obtain the output data after data enhancement.
Optionally, in an embodiment of the present application, the method further includes: and the training module is used for carrying out neural network training by taking the mapping relation as training data, and stopping training when a training termination condition is met to obtain the mapping network.
Optionally, in an embodiment of the application, the enhancing module is specifically configured to input a voltage signal of the input data to the mapped target memristor array, apply random noise through the memristor array, output a current signal of the input data, and convert the current signal to obtain the data-enhanced output data.
Optionally, in an embodiment of the present application, the training termination condition includes: the loss function of the mapping network is smaller than a preset threshold value; and/or the error between the input data and the output data is less than a preset error threshold.
An embodiment of a third aspect of the present application provides an electronic device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor executing the program to perform the memristor noise based data enhancement method as described in the above embodiments.
A fourth aspect of the present application provides a computer-readable storage medium, on which a computer program is stored, where the program is executed by a processor to perform a memristor noise-based data enhancement method as described in the foregoing embodiments.
According to the method and the device, the noise disturbance adding function in data enhancement is achieved through the noise of the memristor, the time required for generating random numbers is avoided, the random Gaussian noise used in the related technology is replaced by the real noise of the memristor in the data enhancement process, data enhancement of limited data can be achieved, the generated image after the data enhancement has diversity and strong randomness, the output data of the memristor is directly input into a task network to be applied, the output data can be deleted after the use, and caching is not needed.
Additional aspects and advantages of the present application will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the present application.
Drawings
The above and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a flow chart of a memristor noise-based data enhancement method provided in accordance with an embodiment of the present application;
FIG. 2 is a schematic diagram of a memristor structure provided according to an embodiment of the present application;
FIG. 3 is a schematic diagram of a memristor array provided in accordance with an embodiment of the present application;
fig. 4 is a schematic diagram of data generation of a generation network according to an embodiment of the present application;
FIG. 5 is a memristor array embedding schematic diagram provided in accordance with an embodiment of the present application;
FIG. 6 is another memristor array embedding schematic provided in accordance with an embodiment of the present application;
FIG. 7 is an example diagram of a memristor noise-based data enhancement device in accordance with an embodiment of the present application;
fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary and intended to be used for explaining the present application and should not be construed as limiting the present application.
The following describes a memristor noise-based data enhancement method, device, electronic device and medium according to embodiments of the present application with reference to the accompanying drawings. Aiming at the problems that an off-line data enhancement mode mentioned in the background technology center is applicable to a small data set, an on-line data enhancement mode is long in time consumption, low in efficiency and single in data enhancement mode, the noise disturbance adding function in data enhancement is achieved through the noise of the memristor, the time needed for generating random numbers is avoided, in the data enhancement process, the real noise of the memristor is used for replacing random Gaussian noise used in the related technology, data enhancement of limited data can be achieved, an image generated after data enhancement has diversity and strong randomness, output data of the memristor is directly input into a task network to be applied, and the output data can be deleted after being used without caching. Therefore, the problems that in the related art, an offline data enhancement mode is applicable to a smaller data set, an online data enhancement mode is long in time consumption and low in efficiency, and a data enhancement mode is single are solved.
Specifically, fig. 1 is a flowchart of a memristor noise-based data enhancement method according to an embodiment of the present application.
As shown in fig. 1, the memristor noise-based data enhancement method includes the following steps:
in step S101, a mapping relationship characterizing a relationship between input data and output data is determined.
It is understood that the mapping relationship in the embodiment of the present application may be an inherent mapping relationship between input data and output data, for example, a known resistance value, an output voltage value according to an input current value and ohm's law, or may be a mapping relationship between input data and output data according to actual requirements, for example, a difference value between an output data value and an input data value is a fixed value or an output data value is a preset multiple of an input data value, and is not limited in particular.
In step S102, based on the mapping relationship, the mapping network corresponding to the mapping relationship is mapped to the target memristor array.
After the mapping relation between the input data and the output data is determined through the step S101, training of the neural network is carried out according to the mapping relation, and the obtained mapping network is mapped to the memristor array.
As shown in fig. 2, a structural form of the memristor is shown, and it can be understood that the memristor itself has noise conforming to gaussian distribution, and the noise fluctuates randomly within a certain range. The noise of the memristor device has the characteristics of uncertainty, randomness and the like, so that the data generated by taking the noise of the memristor device as a disturbance adding source has the characteristics of diversity, randomness and the like. The advantage can effectively avoid the defect brought by data enhancement by a single means, so that the model has better robustness after training.
Optionally, in an embodiment of the present application, the mapping relationship is used as training data to perform neural network training, and when a training termination condition is met, the training is stopped to obtain the mapping network. The input dimension of the mapping network may be equal to the output dimension, or may not be equal to the output dimension.
Optionally, in an embodiment of the present application, the training termination condition includes: the loss function of the mapping network is less than a preset threshold, and/or the error between the input data and the output data is less than a preset error threshold.
The loss function in the embodiment of the present application may be a mean square error or other errors at a pixel level, and is not particularly limited.
Specifically, when the mapping network is trained, the embodiment of the present application stops training when a preset training termination condition is reached, and the training termination condition of the embodiment of the present application may be that a loss function of the mapping network is smaller than a set threshold, for example, when the threshold is a, and when a value of the loss function is smaller than a, the training is stopped. The training termination condition of the embodiment of the present application may also be that an error between the input data and the output data is smaller than a preset error threshold, for example, the input data and the output data are quantized according to a certain standard, a difference between the quantized input data and the quantized output data is compared, and the training is stopped when the difference is smaller than the preset error threshold. The training termination condition of the embodiment of the present application may also be that the training round or the duration reaches a preset value, and the like, for which, a person skilled in the art may set the training termination condition according to the actual situation, and is not specifically limited.
Fig. 2 shows a structure of a single memristor, in an embodiment of the present application, a plurality of memristor structures may be connected, as shown in fig. 3, a plurality of memristors form a memristor array of rows and columns, and a mapping network trained by the above embodiment is mapped to the memristor array of rows and columns.
As shown in fig. 3, x1 to x5 represent original data, and x1 'to x 5' are data obtained after transformation. A mapping network is obtained by training a simple neural network, as shown in fig. 4, it should be noted that fig. 4 only shows the basic principle of training, specific network structures and parameters can be set according to actual situations, and the network has the characteristic that input dimensionality is equal to output dimensionality. The training of the mapping network is trained with the pixel-level mean square error as a loss function, with the goal of reducing the error between the input and output. And mapping the trained mapping network into the memristor array, namely writing network weights into the memristors.
In step S103, input data is input to the mapped target memristor array, and after random noise is applied to the target memristor array, data-enhanced output data is obtained.
It is to be appreciated that the mapped memristor array may output transformed output data from the input data. The input data are input into the task network for operation after passing through the memristor array, the data of the batch can be directly deleted after the use is finished, extra space is not needed for caching, and the defect that a large amount of cache is needed for offline data enhancement is overcome.
Optionally, in an embodiment of the present application, inputting input data to the mapped target memristor array, and obtaining data-enhanced output data after random noise is applied to the target memristor array, includes: and inputting a voltage signal of input data to the mapped target memristor array, applying random noise through the memristor array, outputting a current signal of the input data, and converting the current signal to obtain data-enhanced output data.
Specifically, when data enhancement is performed, a voltage signal of data flows through the memristor array, and a current output signal is obtained. Since random noise exists in the memristor, the obtained output signal can be regarded as data obtained after noise is added to the original data. The noise has the characteristics of strong uncertainty, strong randomness and the like, so that the generated data has the characteristics of diversity, randomness and the like.
The memristor noise-based data enhancement mode of the embodiment of the application can be embedded into any task in an online data enhancement mode, as shown in fig. 5. Before input data are input into the task network, the input data are input into the mapped memristor array, the input data of the current batch are enhanced through noise data, and then the input data are input into the task network, so that the purpose of on-line data enhancement is achieved. The input data flow is directly applied to each network after the array, and then deleted, and the generated data is cached without extra space, so that the defect of offline data enhancement is overcome. Meanwhile, due to the uncertainty of the noise, the noise added to the data input each time is different, so that the defects caused by a single data enhancement means are avoided.
The memristor noise-based data enhancement method in the embodiment of the application can also be used as an independent module, embedded into some special data enhancement networks, and used for adding random disturbance to intermediate codes instead of original data, for example, a self-encoder is shown in fig. 6. The memristor array is embedded into the self-encoder network, replacing the portion of it that would otherwise require the generation of random numbers to add noise. Because the noise of the memristor device is used as an adding source of random disturbance, no additional random number is required to be generated, so that the speed of model training and reasoning is accelerated, and the defects caused by on-line data enhancement are avoided.
By designing the memristor array as an independent data enhancement module and embedding the memristor array into any task, the design scheme is flexible and changeable, and the plug-and-play advantage is achieved.
It should be noted that in the embodiments of the present application, the memristor may be replaced by other devices according to actual needs, such as a Phase Change Memory (PCM) or a ferroelectric memory (FeRAM).
According to the data enhancement method based on the memristor noise, the noise disturbance adding function in data enhancement is achieved through the noise of the memristor, the time needed for generating random numbers is avoided, in the data enhancement process, the random Gaussian noise used in the related technology is replaced by the real noise of the memristor, data enhancement of limited data can be achieved, images generated after data enhancement have diversity and strong randomness, the output data of the memristor are directly input into a task network to be applied, and the images can be deleted after being used without caching. Therefore, the problems that in the related art, an offline data enhancement mode is applicable to a smaller data set, an online data enhancement mode is long in time consumption and low in efficiency, and a data enhancement mode is single are solved.
Next, a data enhancement device based on memristor noise proposed according to an embodiment of the present application is described with reference to the accompanying drawings.
FIG. 7 is an example diagram of a memristor noise-based data enhancement device according to an embodiment of the present application.
As shown in fig. 7, the memristor noise-based data enhancement device 10 includes: an acquisition module 100, a mapping module 200 and an enhancement module 300.
The obtaining module 100 is configured to determine a mapping relationship representing a relationship between input data and output data. The mapping module 200 is configured to map a mapping network corresponding to the mapping relation to the target memristor array based on the mapping relation. The enhancing module 300 is configured to input data to the mapped target memristor array, and obtain data-enhanced output data after random noise is applied to the target memristor array.
Optionally, in an embodiment of the present application, the method further includes: and the training module is used for carrying out neural network training by taking the mapping relation as training data, and stopping training when a training termination condition is met to obtain the mapping network.
Optionally, in an embodiment of the present application, the enhancing module 300 is specifically configured to input a voltage signal of input data to a mapped target memristor array, apply random noise through the memristor array, output a current signal of the input data, and convert the current signal to obtain data-enhanced output data.
Optionally, in an embodiment of the present application, the training termination condition includes: the loss function of the mapping network is less than a preset threshold, and/or the error between the input data and the output data is less than a preset error threshold.
It should be noted that the foregoing explanation of the embodiment of the memristor noise-based data enhancement method is also applicable to a memristor noise-based data enhancement apparatus of the embodiment, and is not repeated here.
According to the data enhancement device based on the memristor noise, the noise disturbance adding function in data enhancement is achieved through the noise of the memristor, the time needed for generating random numbers is avoided, in the data enhancement process, the random Gaussian noise used in the related technology is replaced by the real noise of the memristor, data enhancement can be conducted on limited data, the generated image after the data enhancement has diversity and strong randomness, the output data of the memristor is directly input into a task network to be applied, and the data can be deleted after being used without caching. Therefore, the problems that in the related art, an offline data enhancement mode is applicable to a smaller data set, an online data enhancement mode is long in time consumption and low in efficiency, and a data enhancement mode is single are solved.
Fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present application. The electronic device may include:
a memory 801, a processor 802, and a computer program stored on the memory 801 and executable on the processor 802.
The processor 802, when executing a program, implements a memristor noise-based data enhancement method provided in the above-described embodiments.
Further, the electronic device further includes:
a communication interface 803 for communicating between the memory 801 and the processor 802.
A memory 801 for storing computer programs operable on the processor 802.
The memory 801 may include high-speed RAM memory, and may also include non-volatile memory (non-volatile memory), such as at least one disk memory.
If the memory 801, the processor 802, and the communication interface 803 are implemented independently, the communication interface 803, the memory 801, and the processor 802 may be connected to each other via a bus and perform communication with each other. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended ISA (EISA) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown in FIG. 8, but this is not intended to represent only one bus or type of bus.
Optionally, in a specific implementation, if the memory 801, the processor 802, and the communication interface 803 are integrated on one chip, the memory 801, the processor 802, and the communication interface 803 may complete mutual communication through an internal interface.
The processor 802 may be a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), or one or more Integrated circuits configured to implement embodiments of the present Application.
The present embodiment also provides a computer readable storage medium having stored thereon a computer program, characterized in that the program, when executed by a processor, implements a memristor noise-based data enhancement method as above.
In the description herein, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or N embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the feature. In the description of the present application, "N" means at least two, e.g., two, three, etc., unless explicitly defined otherwise.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more N executable instructions for implementing steps of a custom logic function or process, and alternate implementations are included within the scope of the preferred embodiment of the present application in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the embodiments of the present application.
It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried out in the method of implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and the program, when executed, includes one or a combination of the steps of the method embodiments.
Claims (10)
1. A data enhancement method based on memristor noise is characterized by comprising the following steps:
determining a mapping relation representing the relation between input data and output data;
mapping a mapping network corresponding to the mapping relation to a target memristor array based on the mapping relation; and
inputting the input data to the mapped target memristor array, and obtaining the output data after data enhancement after random noise is applied to the target memristor array.
2. The method of claim 1, further comprising:
and performing neural network training by using the mapping relation as training data, and stopping training when a training termination condition is met to obtain the mapping network.
3. The method of claim 1, wherein the inputting the input data to the mapped target memristor array and obtaining the data-enhanced output data after random noise is applied by the target memristor array comprises:
inputting a voltage signal of the input data to the mapped target memristor array;
outputting a current signal of the input data after applying random noise through the memristor array;
and converting the current signal to obtain the output data after the data enhancement.
4. The method of claim 2, wherein the training termination condition comprises:
the loss function of the mapping network is smaller than a preset threshold value; and/or
An error between the input data and the output data is less than a preset error threshold.
5. A memristor noise-based data enhancement device, comprising:
the acquisition module is used for determining a mapping relation representing the relation between the input data and the output data;
the mapping module is used for mapping a mapping network corresponding to the mapping relation to a target memristor array based on the mapping relation; and
and the enhancing module is used for inputting the input data to the mapped target memristor array and obtaining the output data after data enhancement after random noise is applied to the target memristor array.
6. The apparatus of claim 5, further comprising:
and the training module is used for carrying out neural network training by taking the mapping relation as training data, and stopping training when a training termination condition is met to obtain the mapping network.
7. The apparatus according to claim 5, wherein the enhancement module is specifically configured to input a voltage signal of the input data to the mapped target memristor array, output a current signal of the input data after applying random noise through the memristor array, and convert the current signal to obtain the data-enhanced output data.
8. The apparatus of claim 6, wherein the training termination condition comprises:
the loss function of the mapping network is smaller than a preset threshold value; and/or
An error between the input data and the output data is less than a preset error threshold.
9. An electronic device, comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor executing the program to implement the memristor noise-based data enhancement method as defined in any one of claims 1-4.
10. A computer-readable storage medium having stored thereon a computer program, the program being executable by a processor for implementing a memristor noise-based data enhancement method as defined in any one of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210325808.1A CN114781628A (en) | 2022-03-29 | 2022-03-29 | Memristor noise-based data enhancement method and device, electronic equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210325808.1A CN114781628A (en) | 2022-03-29 | 2022-03-29 | Memristor noise-based data enhancement method and device, electronic equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114781628A true CN114781628A (en) | 2022-07-22 |
Family
ID=82427123
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210325808.1A Pending CN114781628A (en) | 2022-03-29 | 2022-03-29 | Memristor noise-based data enhancement method and device, electronic equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114781628A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109064434A (en) * | 2018-06-28 | 2018-12-21 | 广州视源电子科技股份有限公司 | Image enhancement method and device, storage medium and computer equipment |
| CN109543827A (en) * | 2018-12-02 | 2019-03-29 | 清华大学 | Production fights network equipment and training method |
| CN110956256A (en) * | 2019-12-09 | 2020-04-03 | 清华大学 | Method and device for realizing Bayes neural network by using memristor intrinsic noise |
| US20200193300A1 (en) * | 2018-12-18 | 2020-06-18 | Hewlett Packard Enterprise Development Lp | Systems for introducing memristor random telegraph noise in hopfield neural networks |
| CN112991223A (en) * | 2021-04-06 | 2021-06-18 | 深圳棱镜空间智能科技有限公司 | Image enhancement method, device, equipment and medium based on reversible neural network |
| CN114067157A (en) * | 2021-11-17 | 2022-02-18 | 中国人民解放军国防科技大学 | Memristor-based neural network optimization method and device and memristor array |
| CN114241245A (en) * | 2021-12-23 | 2022-03-25 | 西南大学 | Image classification system based on residual error capsule neural network |
-
2022
- 2022-03-29 CN CN202210325808.1A patent/CN114781628A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109064434A (en) * | 2018-06-28 | 2018-12-21 | 广州视源电子科技股份有限公司 | Image enhancement method and device, storage medium and computer equipment |
| CN109543827A (en) * | 2018-12-02 | 2019-03-29 | 清华大学 | Production fights network equipment and training method |
| US20200193300A1 (en) * | 2018-12-18 | 2020-06-18 | Hewlett Packard Enterprise Development Lp | Systems for introducing memristor random telegraph noise in hopfield neural networks |
| CN110956256A (en) * | 2019-12-09 | 2020-04-03 | 清华大学 | Method and device for realizing Bayes neural network by using memristor intrinsic noise |
| CN112991223A (en) * | 2021-04-06 | 2021-06-18 | 深圳棱镜空间智能科技有限公司 | Image enhancement method, device, equipment and medium based on reversible neural network |
| CN114067157A (en) * | 2021-11-17 | 2022-02-18 | 中国人民解放军国防科技大学 | Memristor-based neural network optimization method and device and memristor array |
| CN114241245A (en) * | 2021-12-23 | 2022-03-25 | 西南大学 | Image classification system based on residual error capsule neural network |
Non-Patent Citations (3)
| Title |
|---|
| HAN ZHAO: "Memristor-based signal processing for edge computing", 《TSINGHUA SCIENCE AND TECHNOLOGY》, 12 November 2021 (2021-11-12), pages 455 * |
| KECHUAN WU: "Enhanced memristor-based MNNs performance on noisy dataset resulting from memristive stochasticity", 《THE INSTITULTION OF ENGINEERING AND TECHNOLOGY》, 1 July 2019 (2019-07-01), pages 704 - 709 * |
| 孙盛阳: "基于忆阻器的卷积神经网络架构研究", 《中国优秀硕士学位论文全文数据库（信息科技辑）》, 15 February 2022 (2022-02-15), pages 135 - 295 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220374688A1 (en) | Training method of neural network based on memristor and training device thereof | |
| CN111652367A (en) | Data processing method and related product | |
| CN111638958B (en) | Cloud host load processing method and device, control equipment and storage medium | |
| US11354238B2 (en) | Method and device for determining memory size | |
| DE102022119386A1 (en) | METHOD AND APPARATUS FOR PERFORMING DENSE PREDICTION USING TRANSFORMER BLOCKS | |
| CN109165730B (en) | State quantization network implementation method in cross array neuromorphic hardware | |
| KR20200099252A (en) | A device for generating verification vector for verifying circuit design, circuit design system including the same and their reinforcement learning method | |
| CN110637306A (en) | Conditional graph execution based on previous reduced graph execution | |
| CN115017178A (en) | Training method and device for data-to-text generation model | |
| CN115879530A (en) | Method for optimizing array structure of RRAM (resistive random access memory) memory computing system | |
| CN114781628A (en) | Memristor noise-based data enhancement method and device, electronic equipment and medium | |
| CN116245141B (en) | Transfer learning architecture, method, electronic device and storage medium | |
| CN115049852B (en) | Bearing fault diagnosis method and device, storage medium and electronic equipment | |
| CN116188896A (en) | Image classification method, system and equipment based on dynamic semi-supervised deep learning | |
| US9336498B2 (en) | Method and apparatus for improving resilience in customized program learning network computational environments | |
| GB2571818A (en) | Selecting encoding options | |
| CN115758905A (en) | Dynamic neural network-based equipment life prediction method and device and electronic equipment | |
| JP7398625B2 (en) | Machine learning devices, information processing methods and programs | |
| US20230075716A1 (en) | Sequence modeling using imputation | |
| CN114118358A (en) | Image processing method, image processing apparatus, electronic device, medium, and program product | |
| CN114169510A (en) | Storage device and operation method thereof | |
| CN118036666B (en) | Task processing method, device, equipment, storage medium and computer program product | |
| CN112882955A (en) | Test case recommendation method and device and electronic equipment | |
| CN111401555A (en) | Model training method, device, server and storage medium | |
| CN116384452B (en) | Dynamic network model construction method, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |