Artificial intelligence system and method for oversampling input data

Abstract

The artificial intelligence system of the present invention may comprise the steps of: calculating importance scores for data points of a dataset; calculating an oversampling rate for each of the data points; calculating sample weights on the basis of the oversampling rate for each of the data points; and oversampling the data points in response to the calculated sample weights.

Description

Artificial intelligence system and method for oversampling input data

[0001] The present invention relates to an artificial intelligence system, and more specifically, to an artificial intelligence system and method for oversampling input data to resolve imbalances in input data and improve the performance of the artificial intelligence system.

[0002] Recently, artificial intelligence (AI) technology has been attracting attention from society as it demonstrates cutting-edge development. Artificial intelligence refers to computers performing unique human intellectual abilities with high proficiency, encompassing concepts such as "a computer brain that executes tasks within the realm of human intelligence," "the engineering and science of creating intelligent machines," and "a system of algorithms designed to think, perceive, and act like humans."

[0003] Artificial intelligence is being introduced as a technology that will provide highly integrated smart spaces when utilized alongside augmented reality, the Internet of Things, edge computing, and digital twins, and is being emphasized as a core new technology that will lead the era of the Fourth Industrial Revolution. Furthermore, AI is attracting attention as a next-generation growth engine capable of evolving industrial ecosystems beyond standardized problem-solving, and is being actively applied not only in IT, healthcare, agriculture, energy, automobiles, and robotics, but also in knowledge service industries such as retail, finance, law, education, real estate, advertising, and telecommunications. In other words, AI is preparing for a new era by combining with all existing systems, ranging from industries that aim to improve convenience or standards in daily life to the entire spectrum of culture and arts in our society.

[0004] In fields such as medicine, finance, and manufacturing, data collection costs are high or rare data is frequently used. In such cases, a small amount of tabular data is used for regression analysis. However, artificial intelligence models trained on such small amounts of data have suffered from low generalization performance, a high risk of overfitting, and performance degradation due to data imbalance and sparsity. To address these issues, data oversampling techniques have been proposed, with representative examples including SMTOE (Synthetic Minority Oversampling Technique) and ADASYN (Adaptive Synthetic Sampling). These conventional data oversampling techniques mitigate data imbalance by generating data from minority classes.

[0005] However, while these conventional oversampling techniques are effective when applied to classification problems where input variables are continuous, they have the problem of being difficult to apply when categorical input variables are included in the data. (Korean Patent Publication No. 10-2024-0105905)

[0006] One embodiment of the present invention aims to provide an artificial intelligence system and method capable of achieving excellent performance even based on small-scale table input data by dynamically selecting an oversampling rate based on the characteristic distribution of input data in the form of table data and performing oversampling.

[0007] A system of one embodiment of the present invention comprises at least one processor; and at least one memory for storing instructions or information executed by the at least one processor; wherein the operation performed by the instructions or information executed by the at least one processor may include an operation of calculating an oversampling rate for each of the data points of a dataset, an operation of calculating a sample weight based on the oversampling rate for the data points, and an operation of oversampling the data points in correspondence with the calculated sample weight.

[0008] Here, the dataset may include multiple data points composed of feature vectors and target values ​​of feature vectors.

[0009] Additionally, the method further includes an operation to calculate importance scores for the above data points, wherein the operation to calculate importance scores may include an operation to calculate an importance score by calculating the change amount of target values ​​of the above data points, and an operation to generate oversampling table data by selecting some of the data points having high importance scores.

[0010] In addition, the oversampling rate can be calculated as the ratio of the number of times each characteristic value of a plurality of characteristics included in the data points is duplicated within the dataset.

[0011] In addition, the sample weight can be calculated by multiplying multiple oversampling rates of the data points.

[0012] In addition, the above data points may be duplicated by the calculated sample weights and included in the dataset.

[0013] A computerized method according to one embodiment of the present invention may include: a step of calculating an oversampling rate for each of the data points of a dataset stored in memory through a processor; a step of calculating a sample weight for each of the data points based on the oversampling rate stored in memory through a processor; and a step of oversampling the data points in correspondence with the calculated sample weight stored in memory through a processor.

[0014] Here, the dataset may be composed of multiple data points consisting of a feature vector and a target value of the feature vector.

[0015] Additionally, the method further includes a step of calculating importance scores for the data points; wherein, in the step of calculating importance scores, the change amount of the target values ​​of the data points is calculated to calculate the importance scores, and some of the data points having high importance scores are selected to generate oversampling table data.

[0016] In addition, the oversampling rate can be calculated as a ratio obtained by dividing the number of times each characteristic value of a plurality of characteristics included in the data points is duplicated within the dataset by the total number of data points.

[0017] In addition, in the step of calculating the sample weight, the sample weight may be calculated by multiplying multiple oversampling rates of the data points and rounding the result.

[0018] In addition, in the step of oversampling the data points, the data points may be duplicated by the calculated sample weights and included in the dataset.

[0019] In a custom integrated circuit comprising a memory in which information and instructions of an embodiment of the present invention are stored and a functional block including at least one processor that requests access to said memory, said memory may store instructions or information including an operation of calculating an oversampling rate for each of the data points of a dataset, calculating a sample weight based on said oversampling rate for said data points, and oversampling said data points in correspondence with said calculated sample weight.

[0020] A computerized method according to one embodiment of the present invention comprises: a step of calculating an oversampling rate for each of the data points of a dataset including tabular cathode material characteristic information stored in memory through a processor; a step of calculating a sample weight based on the oversampling rate for each of the data points stored in memory through a processor; and a step of oversampling the data points in correspondence with the calculated sample weight stored in memory through a processor, wherein at least some of the cathode material characteristic information of the dataset may be oversampled and stored in memory.

[0021] Here, the dataset is composed of a plurality of data points consisting of a cathode material characteristic vector and a target value of the cathode material characteristic vector, and the dataset may include cathode material characteristic information including a cathode active material and doping elements.

[0022] Additionally, the method further includes a step of calculating importance scores for the data points; wherein, in the step of calculating importance scores, the change amount of the target values ​​of the data points is calculated to calculate the importance scores, and some of the data points having high importance scores are selected to generate oversampling table data.

[0023] In addition, the oversampling rate can be calculated as a ratio obtained by dividing the number of times each characteristic value of the doping elements included in the data points is duplicated within the dataset by the total number of data points.

[0024] In addition, in the step of calculating the sample weight, the sample weight may be calculated by multiplying multiple oversampling rates of the data points and rounding the result.

[0025] Additionally, in the step of oversampling the data points, the data points are duplicated by the calculated sample weights and included in the dataset, so that some of the doping elements may be oversampled.

[0026] In addition, the above-mentioned cathode material characteristic information includes aluminum, zirconium, magnesium, titanium, boron, fluorine, tungsten, molybdenum, gallium, nitrogen group, and calcium as doping elements that dope the cathode material, and by calculating the oversampling rate and sample weight of each of the above-mentioned doping elements, the characteristic information of some of the above-mentioned doping elements may be oversampled.

[0027] Additionally, the method further includes the step of inputting data-enhanced cathode material characteristic information into an artificial intelligence model executed by a processor for learning; and the step of predicting the doping conditions of the cathode material using the learned artificial intelligence model executed by the processor; wherein, in the prediction step, the predicted cathode material doping conditions may be derived as at least one of tungsten / zirconium (W / Zr), tungsten / aluminum (W / Al), or tungsten / titanium (W / Ti).

[0028] A computing system comprising a user computing device of an embodiment of the present invention and a server computing system that performs a task corresponding to instructions or data received from the user computing device, wherein the server computing system comprises: a database that stores a dataset for training an artificial intelligence model; a memory that stores instructions and data for training the artificial intelligence model and performing the task; and at least one processor that performs the task according to the instructions or data of the memory, wherein the at least one processor may perform an operation of receiving an oversampling instruction for the dataset from the user computing device, calculating an oversampling rate for each of the data points of the dataset, calculating a sample weight based on the oversampling rate for the data points, oversampling the data points in correspondence with the calculated sample weight, and providing the oversampling result to the user computing device.

[0029] Here, the at least one processor may further perform the operation of inputting an oversampled dataset into an artificial intelligence model executed by the processor for training, executing a prediction task based on instructions or data received from the user computing device, and providing the prediction task result to the user computing device for output.

[0030] In addition, the above dataset may include multiple data points composed of feature vectors and target values ​​of feature vectors.

[0031] Additionally, the above at least one processor further performs an operation to calculate an importance score for the data points, and the operation to calculate the importance score may include an operation to calculate an importance score by calculating the change amount of the target values ​​of the data points, and an operation to generate oversampling table data by selecting some of the data points having a high importance score.

[0032] In addition, the oversampling rate can be calculated as a ratio obtained by dividing the number of times each characteristic value of a plurality of characteristics included in the data points is duplicated within the dataset by the total number of data points.

[0033] In addition, the sample weight can be calculated by multiplying multiple oversampling rates of the data points and rounding the result.

[0034] In addition, the above data points may be duplicated by the calculated sample weights and included in the dataset.

[0035] According to one embodiment of the present invention, the performance of an artificial intelligence system can be improved even based on small-scale table input data by dynamically determining the oversampling rate based on the characteristic distribution of the input data.

[0036] FIG. 1 is a schematic diagram of an electronic device according to an embodiment of the present invention.

[0037] FIG. 2 is a schematic diagram illustrating an oversampling method according to an embodiment of the present invention.

[0038] FIG. 3 is a schematic diagram showing an input dataset of an embodiment of the present invention.

[0039] FIG. 4 is a schematic diagram illustrating an oversampling process according to an embodiment of the present invention.

[0040] FIGS. 5 and 6 are schematic diagrams showing the results of evaluating an oversampled input dataset according to an embodiment of the present invention.

[0041] To clarify the technical concept of the present disclosure, embodiments of the present invention will be described in detail with reference to the attached drawings. In describing the present disclosure, detailed descriptions of related known functions or components will be omitted if it is determined that such detailed descriptions would unnecessarily obscure the essence of the present disclosure. Components having substantially the same functional configuration in the drawings have been assigned the same reference numerals and symbols as much as possible, even if they are shown in different drawings. For convenience of explanation, devices and methods will be described together where necessary. Each operation of the present disclosure does not necessarily need to be performed in the order described and may be performed in parallel, selectively, or individually.

[0042] The terms used in the embodiments of this disclosure have been selected to be as widely used and general as possible, taking into account the functions of this disclosure; however, these terms may vary depending on the intent of those skilled in the art, case law, the emergence of new technologies, etc. Additionally, in specific cases, terms have been selected at the applicant's discretion, and in such cases, their meanings will be described in detail in the description of the relevant embodiments. Therefore, terms used in this specification should be defined not merely by their names, but based on their meanings and the overall content of this disclosure.

[0043] Throughout this disclosure, singular expressions may include plural expressions unless the context clearly indicates otherwise. Terms such as “comprising” or “having” are intended to specify the presence of features, numbers, steps, actions, components, parts, or combinations thereof, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof. That is, throughout this disclosure, when a part is described as “comprising” a certain component, it means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0044] Expressions such as "at least one" modify the entire list of components and do not modify the components of the list individually. For example, "at least one of A, B, and C" and "at least one of A, B, or C" refer to only A, only B, only C, both A and B, both B and C, both A and C, all of A, B, and C, or any combination thereof.

[0045] Additionally, terms such as “...part,” “...module,” etc., as described in this disclosure refer to a unit that processes at least one function or operation, and may be implemented in hardware or software, or a combination of hardware and software.

[0046] Throughout the entire disclosure, when a part is described as being “connected” to another part, this includes not only cases where they are “directly connected” but also cases where they are “electrically connected” with other elements interposed between them. Furthermore, when a part is described as “comprising” a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0047] As used throughout this disclosure, the expression “configured to” may be replaced, depending on the context, with, for example, “suitable for,” “having the capacity to,” “designed to,” “adapted to,” “made to,” or “capable of.” The term “configured to” may not necessarily mean only “specifically designed to” in hardware. Instead, in some situations, the expression “system configured to” may mean that the system is “capable of” together with other devices or components. For example, the phrase “a processor configured (or set) to perform A, B, and C” may mean a dedicated processor for performing said operations (e.g., an embedded processor), or a generic-purpose processor (e.g., a CPU or an application processor) capable of performing said operations by executing one or more software programs stored in memory.

[0048] Artificial intelligence (AI) is a field of computer science and information technology that studies methods to enable computers to perform thinking, learning, and self-development—tasks achievable by human intelligence—and refers to the ability of computers to mimic intelligent human behavior. Furthermore, AI does not exist in isolation but is closely related, directly or indirectly, to many other fields of computer science. Particularly in the modern era, there are very active attempts to introduce AI elements into various sectors of information technology and utilize them to solve problems within those fields.

[0049] Machine learning is a field of artificial intelligence that enables computers to learn without explicit programming. Specifically, machine learning can be defined as a technology that studies and builds systems and algorithms capable of learning, making predictions, and improving their own performance based on empirical data. Rather than executing strictly defined static program commands, machine learning algorithms adopt an approach of constructing specific models to derive predictions or decisions based on input data. The term 'machine learning' may be used interchangeably with 'machine learning'.

[0050] Many machine learning algorithms have been developed to address how to classify data in machine learning. Representative examples include Decision Trees, Bayesian Networks, Support Vector Machines (SVMs), and Artificial Neural Networks (ANNs). A Decision Tree is an analytical method that performs classification and prediction by plotting decision rules in a tree structure. A Bayesian Network is a model that represents the probabilistic relationships (conditional independence) between multiple variables in a graph structure. Bayesian Networks are suitable for data mining through unsupervised learning. Support Vector Machines are supervised learning models for pattern recognition and data analysis, primarily used for classification and regression analysis. Artificial Neural Networks model the operating principles of biological neurons and the relationships between them; they are information processing systems in which multiple neurons, referred to as nodes or processing elements, are connected in a layered structure.

[0051] Artificial neural networks are models used in machine learning, serving as statistical learning algorithms in machine learning and cognitive science that draw inspiration from biological neural networks (particularly the brain within the animal central nervous system). Specifically, an artificial neural network can refer to a model in which artificial neurons (nodes), forming a network through synaptic connections, change the strength of these connections through learning to possess problem-solving capabilities. The term artificial neural network may be used interchangeably with neural network.

[0052] An artificial neural network may include multiple layers, and each layer may include multiple neurons. Additionally, an artificial neural network may include synapses connecting neurons. An artificial neural network can generally be defined by the following three factors: ㉠ connection patterns between neurons of different layers, ㉡ a learning process that updates the weights of the connections, and ㉢ an activation function that generates an output value from a weighted sum of inputs received from the previous layer.

[0053] Artificial neural networks may include, but are not limited to, network models such as Deep Neural Networks (DNN), Recurrent Neural Networks (RNN), Bidirectional Recurrent Deep Neural Networks (BRDNN), Multilayer Perceptrons (MLP), and Convolutional Neural Networks (CNN). In this specification, the term 'layer' may be used interchangeably with the term 'layer'.

[0054] Artificial neural networks are classified into single-layer neural networks and multi-layer neural networks depending on the number of layers. A typical single-layer neural network consists of an input layer and an output layer. Additionally, a typical multi-layer neural network consists of an input layer, one or more hidden layers, and an output layer.

[0055] The input layer is a layer that receives external data, and the number of neurons in the input layer is equal to the number of input variables. The hidden layer is located between the input layer and the output layer, receives signals from the input layer, extracts features, and transmits them to the output layer. The output layer receives signals from the hidden layer and outputs an output value based on the received signals. Input signals between neurons are multiplied by their respective connection strengths (weights) and then summed; if this sum is greater than the neuron's threshold, the neuron is activated and outputs the value obtained through the activation function.

[0056] Meanwhile, a deep neural network containing multiple hidden layers between the input layer and the output layer can be a representative artificial neural network that implements deep learning, a type of machine learning technique. Meanwhile, the term 'deep learning' may be used interchangeably with the term 'deep learning,' and the term 'learning' may be used interchangeably with 'training.'

[0057] The machine learning workflow consists of a series of processes involving collecting data for learning and validation, modeling, and training the model, and may include the processes of collecting training data, checking and exploring data, data preprocessing and cleaning, modeling, and training.

[0058] 1. Collect Training Data

[0059] Training data applied to the training of the learning model of this specification may be generated using data collected from a plurality of samples. In this specification, at least one different type of training data set may be used to train the learning model, and each training data may further include one or more experimental results used as feature labels. At least a portion of the training data set may be used to train the learning model, and another portion may be used to validate the learned learning model.

[0060] 2. Data Inspection and Exploration

[0061] Once training data for training a learning model is collected, the collected training data can be examined and explored regarding its structure, noise data, and data cleaning methods for machine learning applications.

[0062] This stage of data inspection and exploration is called Exploratory Data Analysis (EDA), which can be described as the process of observing and understanding collected data from various angles. Before training the data, independent variables, dependent variables, variable types, and data types are examined using visualizations such as graphs and statistical tests, allowing the characteristics of the data and inherent structural relationships to be identified in advance. Through this EDA, examining the distribution and values ​​of the data enables a better understanding of the phenomena represented by the data and the discovery of potential problems. Furthermore, by examining the data from various angles, diverse patterns that might not have been identified during the problem definition stage can be discovered, allowing for the modification of existing hypotheses or the formulation of new ones. Exploratory data analysis can broadly encompass the process of searching for data outliers and analyzing the relationships between data attributes.

[0063] The process of detecting outliers involves verifying whether the data contains them and can include sampling, statistical, and visualization methods. Sampling methods involve drawing random samples from the data to identify overall trends and anomalies in the data values. Statistical methods may utilize summary statistics, such as the mean, median, and mode to identify the center of the data, or range and variance to check the dispersion. Visualization methods utilize probability density functions, histograms, dot plots, word clouds, time series charts, and maps to determine which statistical indicators are appropriate for the individual attributes of the collected data. However, when using statistical indicators, caution should be exercised regarding the use of statistical indicators: while the mean reflects all data values ​​within a set and is therefore affected by outliers, the median uses only the single value in the middle, allowing for representative results even in the presence of outliers.

[0064] The process of analyzing relationships between data attributes involves identifying combinations of attributes within the data that possess meaningful correlations. Relationship analysis can be conducted differently depending on the combination of attributes between qualitative attributes (Categorical Variables; Qualitative), which cannot be expressed numerically but can be arbitrarily quantified, and quantitative attributes (Numeric Variables; Quantitative), which can be quantified. Categorical-categorical relationships can display the number of values ​​corresponding to each pair of attribute values ​​using cross-tabulation tables or mosaic plots; Numeric-categorical relationships can be visually represented through box plots or by observing statistical values ​​by category (mean, median, etc.); and Numeric-numeric relationships can analyze the association between two attributes using correlation coefficients. It can be confirmed that a correlation coefficient of -1 indicates a negative correlation where the two attributes change in opposite directions, 0 indicates no correlation, and 1 indicates a positive correlation where the two attributes always change in the same direction. The relationship between two attributes with a correlation coefficient can also exhibit various aspects, which can be visually represented using a scatter plot.

[0065] 3. Data Preprocessing and Cleansing

[0066] Data that has completed inspection and exploration undergoes data preprocessing to transform it into a format suitable for machine learning training models. Data preprocessing involves cleaning the data and converting it into a form that the model can understand; it generally includes handling missing data, outlier removal, data scaling, categorical data encoding, feature selection and extraction, and data transformation. The detailed processes of data preprocessing may be performed in whole or in part selectively, and a separate machine learning model may be used for this purpose.

[0067] Handling Missing Data is the process of handling missing values ​​when they exist in the data; these values ​​can be displayed as NaN (Not a Number) or empty, or deleted. Filling in or deleting missing values ​​improves data completeness, and values ​​such as the mean, median, or mode may be used when filling in missing values.

[0068] Outlier removal is the process of eliminating outliers, which are values ​​that deviate from typical data patterns. Since outliers can degrade model performance, they must be removed or replaced; this involves identifying outliers and deleting the corresponding rows or columns or replacing them with other values.

[0069] Data scaling is the process of adjusting the size of data; through data scaling, the range of the data is adjusted, which can improve model performance or accelerate convergence. Data scaling allows data characteristics to be aligned within a similar range, and generally, standardization and normalization can be applied. Standardization is a method of transforming data into a distribution with a mean of 0 and a standard deviation of 1; it is primarily performed using the mean and standard deviation, and the standardized value z is It can be denoted as (where x is the original value, μ is the mean, and σ is the standard deviation). Normalization is a method of transforming the range of data to [0,1] or [-1,1], primarily using minimum and maximum values ​​to transform the data, and the normalized value x norm silver It can be expressed as (x is the original value, x min is the minimum value, x max is the maximum value).

[0070] Categorical Data Encoding is the process of converting categorical variables, which are represented as string or integer values ​​and cannot be directly input into a model, into a numeric type that can be input. Generally, one-hot encoding or label encoding can be used to convert categorical variables into numeric types.

[0071] Feature selection and extraction is intended to improve the performance of a model by selecting the most useful features for model training or extracting new features. Through this process, the complexity of the model can be reduced and overfitting can be prevented.

[0072] Data transformation involves converting data to extract new information or enable a model to understand it better, and may include the tokenization of text data or the preprocessing of image data. Through data transformation, model performance can be improved by extracting useful features from original data or converting data into an appropriate format.

[0073] Through data preprocessing as described above, it is possible to achieve the effects of improving the performance and ensuring the stability of machine learning models.

[0074] Meanwhile, when training a learning model according to one embodiment of the present invention, a process of preprocessing information written in natural language and a process of training a language model based on the preprocessed data may be performed.

[0075] 3-1. Text Preprocessing for Large-Scale Language Models

[0076] If the collected data has not been preprocessed according to the requirements, tokenization, cleaning, and normalization can be performed to suit the intended use of the data.

[0077] Tokenization refers to the process of dividing given data into units called tokens, and these token units can generally be defined as meaningful units. Tokenization can broadly include word tokenization and sentence tokenization.

[0078] Tokenization refers to the process of dividing given data into units called tokens, and these token units can generally be defined as meaningful units. Tokenization can broadly include word tokenization and sentence tokenization.

[0079] Word tokenization refers to the case where the standard for tokens is a word; here, a word can include not only individual words but also phrases or meaningful strings. Word tokenization means separating words based on spaces or punctuation marks, such as periods, commas, question marks, semicolons, and exclamation marks. However, since removing all punctuation or special characters during the tokenization process can cause tokens to lose their meaning, precise algorithms may be required. For instance, if a word itself contains punctuation or uses special characters with meaning, simply removing them may not be sufficient. Therefore, tokenization rules such as Penn Treebank Tokenization rules may be applied during the process.

[0080] Sentence tokenization refers to the process of dividing text into sentence units. Typically, if data is unrefined, the corpus is not organized into sentences, so sentence tokenization may be necessary to suit the intended use. Various rules for this sentence tokenization can be defined depending on the language used and how special characters are utilized within the corpus.

[0081] The process of classifying tokens according to their purpose is called tokenization, and before and after tokenization, text data undergoes cleaning and normalization tailored to its intended use. Cleaning involves removing noise data, while normalization involves consolidating words with different representations into a single word.

[0082] Cleansing is sometimes performed prior to tokenization to exclude elements that interfere with the process, but it can also be repeatedly carried out after tokenization to remove noise that remains. The noise data removed during cleansing consists of meaningless characters; methods for eliminating unnecessary words include stopword removal, as well as removing infrequent and short words.

[0083] Normalization tasks include the consolidation of words with different spellings and case consolidation based on rules. Case consolidation is a normalization method that can reduce the number of words in English-speaking languages; since uppercase letters are used only in specific situations, such as at the beginning of a sentence, and most text is written in lowercase, case consolidation can mostly be accomplished through the conversion of uppercase letters to lowercase.

[0084] To process natural language in computing systems, a preprocessing step of converting text into numerical values ​​is required; for this purpose, each word in the text is mapped to a unique integer. This mapping process can utilize techniques such as integer encoding, padding, and one-hot encoding.

[0085] Integer encoding is a method of assigning integers to words. It involves creating a vocabulary sorted by frequency and assigning integers sequentially from lowest to highest frequency. Integer encoding performs sentence tokenization on text data containing multiple sentences, and simultaneously conducts word tokenization through cleaning and normalization processes. During this process, words are converted to lowercase to standardize the word count, and words can be removed based on stop words or word length. Through this, words can be recorded as keys and their frequencies as values. Integer encoding is performed by sorting words within the text in order of frequency and assigning integers to the words with the highest frequencies.

[0086] Padding is a process used to arbitrarily equalize the lengths of sentences of different lengths within a text. Computing systems can perform parallel operations by grouping sentences of equal length into a single matrix. In other words, to perform parallel operations, the lengths of sentences can be equalized by arbitrarily filling the integer encoding results of sentences of different lengths within the text with '0's. That is, the longest sentence is identified from the set of integer-encoded words, and "0"s can be added to the integer matrix corresponding to the length of that longest sentence. The computing system can proceed with parallel processing by recognizing sentences of equal length as a single matrix, and in this process, the "0" words, which are perceived as meaningless, can be ignored. Adjusting the size (shape) of data by filling it with specific values ​​in this manner is called padding, and when the number "0" is used to adjust length, it is referred to as zero padding.

[0087] One-hot encoding is a vector representation method in which the size of the word set is used as the dimension of the vector, and a value of 1 is assigned to the index of the word to be represented and 0 to other indices; the vector represented in this way is called a one-hot vector. One-hot encoding consists of integer encoding and index assignment processes. After integer encoding is performed to assign a unique integer to each word, the unique integer of the word to be represented is considered as the index, and a "1" is assigned to that position, while a "0" is assigned to the index positions of other words. However, one-hot encoding has the disadvantage that the space required to store the vector increases (increase in vector dimension) as the number of words increases, and it is also impossible to verify similarity between words. To address these drawbacks, techniques that vectorize in a multi-dimensional space by reflecting the latent meaning of words are available. These include count-based vectorization methods such as LSA (Latent Semantic Analysis); prediction-based vectorization methods such as NNLM, RNNLM, Word2Vec, and FastText; and the GloVe method, which uses both count-based and prediction-based approaches.

[0088] Meanwhile, in order for a computer to understand and process text, it must be appropriately converted into numbers. Since the performance of natural language processing varies significantly depending on how words are represented, many techniques have been proposed to quantify words. Currently, word embedding, which vectorizes each word through artificial neural network learning, is the most widely used method.

[0089] Word embedding is a method of representing words as vectors, converting words into dense representations. The result derived through the word embedding process is called a dense vector or embedding vector. Word embedding methodologies include LSA, Word2Vec, FastText, and Glove.

[0090] 4. Modeling and Training

[0091] Artificial neural networks can be trained using training data. Here, training refers to the process of determining the parameters of an artificial neural network using training data to achieve objectives such as classifying, regressing, or clustering input data. Typical examples of artificial neural network parameters include weights assigned to synapses or biases applied to neurons.

[0092] An artificial neural network trained on training data can classify or cluster input data according to the patterns of the input data. Meanwhile, an artificial neural network trained using training data may be referred to as a trained model in this specification.

[0093] The following explains the learning methods of artificial neural networks. The learning methods of artificial neural networks can be broadly classified into supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning.

[0094] Supervised learning is a method of machine learning designed to infer a function from training data. Among the functions inferred in this way, outputting a continuous value is called regression, and predicting and outputting the class of an input vector is called classification.

[0095] In supervised learning, an artificial neural network is trained with labels for the training data. Here, a label refers to the correct answer (or result value) that the artificial neural network must infer when training data is input into the artificial neural network. In this specification, the correct answer (or result value) that the artificial neural network must infer when training data is input is referred to as a label or labeling data. Furthermore, in this specification, setting labels on the training data for the training of the artificial neural network is referred to as labeling the training data. In this case, the training data and the corresponding labels constitute a single training set, and can be input to the artificial neural network in the form of a training set.

[0096] Meanwhile, training data represents multiple features, and labeling the training data implies that labels are attached to the features represented by the training data. In this case, the training data can represent the features of the input object in the form of a vector. An artificial neural network can infer a function regarding the association between the training data and the labeled data by utilizing the training data and the labeled data. Furthermore, the parameters of the artificial neural network can be determined (optimized) through the evaluation of the function inferred by the network.

[0097] Unsupervised learning is a type of machine learning in which no labels are provided for the training data. Specifically, unsupervised learning may be a learning method in which an artificial neural network is trained to find and classify patterns within the training data itself, rather than the relationship between the training data and the corresponding labels. Examples of unsupervised learning include clustering or Independent Component Analysis. In this specification, the term 'clustering' may be used interchangeably with the term 'clustering'.

[0098] Examples of artificial neural networks that utilize unsupervised learning include Generative Adversarial Networks (GANs) and Autoencoders (AEs).

[0099] Generative Adversarial Networks (GANs) are machine learning methods in which two distinct artificial intelligence models—a generator and a discriminator—compete to improve performance. In this context, the generator is a model that creates new data, capable of generating new data based on original data. The discriminator, on the other hand, is a model that recognizes data patterns, performing the role of distinguishing whether input data is original data or new data generated by the generator. Furthermore, the generator learns by receiving input data that failed to deceive the discriminator, while the discriminator learns by receiving input data that was deceived by the generator. Consequently, the generator can evolve to deceive the discriminator as effectively as possible, and the discriminator can evolve to better distinguish between original data and data generated by the generator.

[0100] An autoencoder is a neural network that aims to reproduce the input itself as the output. An autoencoder includes an input layer, at least one hidden layer, and an output layer. In this case, since the number of nodes in the hidden layer is less than the number of nodes in the input layer, the dimensionality of the data is reduced, and accordingly, compression or encoding is performed. Additionally, the data output from the hidden layer enters the output layer. In this case, since the number of nodes in the output layer is greater than the number of nodes in the hidden layer, the dimensionality of the data is increased, and accordingly, decompression or decoding is performed.

[0101] Meanwhile, an autoencoder represents input data as hidden layer data by adjusting the connection strengths of neurons through learning. In the hidden layer, information is represented with fewer neurons than in the input layer, and the fact that input data can be reproduced as output implies that the hidden layer has discovered and represented hidden patterns from the input data.

[0102] Semi-supervised learning is a type of machine learning that refers to a learning method utilizing both labeled and unlabeled training data. One technique within semi-supervised learning involves inferring labels from unlabeled training data and then performing learning using those inferred labels; this method can be particularly useful when the cost of labeling is high.

[0103] Reinforcement learning is a theory that states that if an agent is provided with an environment where it can determine the best action to take at every moment, it can find the optimal path through experience alone, without relying on data. Reinforcement learning is primarily executed via a Markov Decision Process (MDP). To explain the MDP, first, an environment is provided containing the information necessary for the agent to take its next action; second, the agent's behavior within that environment is defined; third, rewards are determined for success and penalties for failure; and fourth, the optimal policy is derived through repeated experience until future rewards reach their peak.

[0104] The structure of an artificial neural network is determined by the configuration of the model, activation function, loss function or cost function, learning algorithm, optimization algorithm, etc., and hyperparameters are set in advance before learning, and model parameters are set through learning thereafter, so the content can be determined.

[0105] For example, factors determining the structure of an artificial neural network may include the number of hidden layers, the number of hidden nodes included in each hidden layer, the input feature vector, the target feature vector, etc.

[0106] Hyperparameters include various parameters that must be initially set for training, such as the initial values ​​of model parameters. Model parameters, on the other hand, include various parameters intended to be determined through training. For example, hyperparameters may include initial values ​​for inter-node weights, initial values ​​for inter-node bias, mini-batch size, number of training iterations, and learning rate. Additionally, model parameters may include inter-node weights and inter-node bias.

[0107] A loss function can be used as an indicator (criterion) to determine optimal model parameters during the learning process of an artificial neural network. In an artificial neural network, learning refers to the process of manipulating model parameters to reduce the loss function, and the objective of learning can be viewed as determining model parameters that minimize the loss function. The loss function can primarily be the Mean Squared Error (MSE) or the Cross Entropy Error (CEE), but the present invention is not limited thereto. The Cross Entropy Error can be used when the correct label is one-hot encoded. One-hot encoding is an encoding method in which the correct label value is set to 1 only for neurons corresponding to the correct answer, and the correct label value is set to 0 for neurons that are not the correct answer.

[0108] In machine learning or deep learning, learning optimization algorithms can be used to minimize the loss function, and learning optimization algorithms include Gradient Descent (GD), Stochastic Gradient Descent (SGD), Momentum, Nesterov Accelerate Gradient (NAG), Adagrad, AdaDelta, RMSProp, Adam, Nadam, etc.

[0109] Gradient Descent is a technique that adjusts model parameters in a direction that reduces the loss function value by considering the gradient of the loss function from the current state. The direction in which model parameters are adjusted is called the step direction, and the magnitude of the adjustment is called the step size. In this context, the step size can refer to the learning rate. Gradient Descent obtains the gradient by taking the partial derivative of the loss function with respect to each model parameter, and updates the model parameters by changing them in the direction of the obtained gradient by the learning rate.

[0110] Stochastic Gradient Descent is a technique that divides training data into mini-batches and performs gradient descent on each mini-batch to increase the frequency of gradient descent.

[0111] Adagrad, AdaDelta, and RMSProp are techniques that improve optimization accuracy in SGD by adjusting the step size. In SGD, Momentum and NAG are techniques that improve optimization accuracy by adjusting the step direction. Adam is a technique that improves optimization accuracy by combining Momentum and RMSProp to adjust both the step size and the step direction. Nadam is a technique that improves optimization accuracy by combining NAG and RMSProp to adjust both the step size and the step direction.

[0112] The learning speed and accuracy of artificial neural networks are characterized by being heavily dependent on hyperparameters, as well as the network structure and the type of learning optimization algorithm. Therefore, to obtain a good learning model, it is important to set appropriate hyperparameters in addition to determining a suitable network structure and learning algorithm.

[0113] Typically, hyperparameters are experimentally set to various values ​​while training the artificial neural network, and then set to the optimal value that provides stable training speed and accuracy based on the training results.

[0114] Embodiments according to the artificial intelligence system and method for oversampling input data of an embodiment of the present invention can be applied to the field of developing new materials. In particular, this embodiment can be applied to the development of manganese-rich (Mn-Rich) cathode materials that utilize more than 60% of manganese (Mn), which is a relatively inexpensive material that has a structure different from existing cathode materials applied to batteries. In addition, the embodiments of the present invention can be applied to the field of predicting the structure and physical properties of other structures, such as crystal structures, molecular structures, protein structures, catalyst structures, and metal-organic frameworks (MOFs). Furthermore, an embodiment of the present invention can predict the EC number of a protein having an input amino acid sequence from protein amino acid sequence information input by a user with improved performance, and can predict the EC number of a protein having an input amino acid sequence with improved performance compared to conventional technology, not only at low levels but also at high levels.

[0115] FIG. 1 is a schematic diagram of an electronic device according to one embodiment of the present invention.

[0116] As illustrated in FIG. 1, an electronic device (100) according to one embodiment of the present invention may include at least one processor (110), a memory (120), and a communication unit (130). The electronic device (100) is a basic configuration for performing a computing environment, and in other embodiments, the electronic device (100) may be implemented by including some other components additionally or substantially, implemented as a single or multiple entity, or implemented as only some of the disclosed configurations. Internal or external components of the electronic device (100), or at least some of the components, may transmit or receive data or signals by being connected to each other through a BUS, GPIO (General Purpose Input / Output), SPI (Serial Peripheral Interface), or MIPI (Mobile Industry Processor Interface), etc.

[0117] The processor (110) may mean a set of one or more processors unless the context clearly indicates otherwise, and can control components of the processor (110) and the electronic device (100) by running software (e.g., instructions, programs, etc.) stored in memory (120). Additionally, the processor (110) can perform various operations such as computation, processing, data generation or processing, and can read data from memory (120) or store it in memory (120). The processor (110) may be composed of at least one core and may include a processor for data analysis, machine learning (ML), or deep learning (DL), such as a Central Processing Unit (CPU), a General Purpose Graphics Processing Unit (GPGPU), or a Tensor Processing Unit (TPU). The processor (110) can read software stored in memory (120) and perform data processing for machine learning (or deep learning) according to the present invention. According to one embodiment of the present disclosure, the processor (110) can perform operations for training a neural network. The processor (110) can perform operations for training a neural network, such as processing input data for training in deep learning, extracting features from input data, calculating errors, and updating the weights of the neural network using backpropagation. At least one of the CPU, GPGPU, and TPU of the processor (110) can process the training of the neural network model. For example, the CPU and GPGPU can together process the training of the neural network model and data classification using the neural network model. In addition, in one embodiment of the present disclosure, at least one processor (110) of the electronic device (100) can be used together to process the training of the neural network model and data classification using the neural network model.

[0118] Memory (120) is intended to store various data, and the data may include software (e.g., instructions, programs, etc.) which is acquired, processed, or used by at least one component of the electronic device (100). Unless otherwise clearly expressed in the context, memory (120) may mean a set of one or more memories and may include at least one type of storage medium among flash memory type, hard disk type, multimedia card micro type, card type memory (e.g., SD or XD memory, etc.), RAM, SRAM (Static Random Access Memory), ROM, EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory), magnetic memory, magnetic disk, optical disk, and web storage that performs storage functions on the internet. The instruction, program, or software stored in the memory (120) may be used to refer to an operating system, an application, or middleware that provides various functions to an application to enable the application to utilize the components of the electronic device (100) for controlling the components of the electronic device (100). In one embodiment, when the processor (110) performs a specific operation, the memory (120) may store instructions that are performed by the processor (110) and correspond to the specific operation.

[0119] The communication unit (130) performs wireless or wired communication between the electronic device (100) and another device (e.g., a user terminal or another server), and the communication unit (130) may use wireless communication systems according to methods such as eMBB, URLLC, MMTC, LTE, LTE-A, NR, UMTS, GSM, CDMA, WCDMA, TDMA, FDMA, OFDMA, SCFDMA, WiBro, WiFi, Bluetooth, NFC, GPS, or GNSS. In addition, the communication unit (130) can use various wired communication systems such as USB, HDMI, RS-232 (Recommended Standard-232), POTS (Plain Old Telephone Service), Public Switched Telephone Network (PSTN), xDSL (x Digital Subscriber Line), RADSL (Rate Adaptive DSL), MDSL (Multi Rate DSL), VDSL (Very High Speed ​​DSL), UADSL (Universal Asymmetric DSL), HDSL (High Bit Rate DSL), and Local Area Network (LAN). In one embodiment of the present invention, the communication unit (130) can be configured regardless of the mode of communication, such as wired or wireless, and can be configured with various communication networks such as a Personal Area Network (PAN) and a Wide Area Network (WAN). In addition, the communication network may be the known World Wide Web (WWW), and may also utilize wireless transmission technologies used for short-range communication, such as Infrared Data Association (IrDA) or Bluetooth.

[0120] An electronic device (100) according to an embodiment of the present invention can execute software that configures an artificial intelligence system or a method for oversampling input data.

[0121] As illustrated in FIG. 2, an oversampling method of an artificial intelligence system according to one embodiment of the present invention may include the steps of calculating importance scores for data points of a dataset (S110), calculating an oversampling rate for each of the data points (S120), calculating sample weights based on the oversampling rates for the data points (S130), and oversampling the data points in correspondence with the calculated sample weights (S140).

[0122] The tabular data of one embodiment is structured data composed of rows and columns, where each row represents a sample and each column represents a characteristic. When the tabular data has n rows (samples) and m columns (characteristics), the samples are s n The characteristic is f m It can be written as.

[0123] In the step (S110) of calculating importance scores for data points of a dataset, an artificial intelligence system of one embodiment of the present invention may obtain tabular data as an input dataset. An artificial intelligence system of one embodiment particularly targets small-scale tabular data including categorical inputs. When an input dataset such as tabular data is denoted as D, D is a set of multiple data points, and, here x i is as a feature vector (a set of p-dimensional real vectors) and y i is the target value (set of real numbers), where N is the total number of samples (data points) in the input dataset. The system and method of one embodiment may select d important features and apply oversampling based on the frequency of the feature values. In one embodiment, x of a single data point i f i Let f be the characteristic of yi. j It can be said.

[0124] Each characteristic f of multiple data points j Regarding, prediction change This is calculated, and through it, the importance of each characteristic can be evaluated. The importance score can be calculated as the average of the predicted changes across all samples (data points). The importance score can be calculated using the following mathematical formula.

[0125]

[0126] Based on importance scores, multiple important characteristics with the highest d importance scores can be selected, and the set of selected important characteristics A set of features, i.e., oversampling table data, can be generated by performing oversampling based on data points having the top d important features. With reference to FIG. 3, oversampling table data can be generated in which oversampling is performed by data points including A, B, and C, which are three (d=3) of the important features, in feature f2.

[0127] [Algorithm]

[0128] Step 1: Feature Importance Calculation

[0129] Perturb and compute prediction changes

[0130] Calculate Importance = =

[0131] Select top d important features to form the set

[0132] In the step (S120) of calculating the oversampling rate for each of the data points, each characteristic of set F, i.e., each characteristic f of the oversampling table data i , f j Regarding this, it can be calculated as the frequency of the characteristic value x for each feature, that is, the ratio of the number of times the characteristic value is duplicated within the data. Frequency of characteristic value x is feature f within the dataset i It can be calculated as a ratio, and characteristic f i The frequency of can be equal to the mathematical formula below.

[0133]

[0134] Referring to Figure 3, the dataset selected based on importance scores, i.e., the oversampling table data, contains 10 data points (10 rows), and the feature value "3" of feature f1 contains 5 of the 10 data points, so the frequency p can be 0.5. The "C" of feature f2 contains 1 of the 10 data points, so the frequency p can be 0.1.

[0135] The oversampling rate r for each characteristic value can be defined as shown in the following mathematical formula.

[0136]

[0137] [Algorithm]

[0138] Step 2: Feature Frequency Rate Calculation

[0139] for each feature in do

[0140] for each unique value x of do

[0141] Compute the frequency

[0142] Calculate the oversampling rate

[0143] end for

[0144] end for

[0145] Referring to FIG. 3, the "1" of characteristic f1 yields an oversampling rate of 0.8, and then 1 is added so that the oversampling rate (r1) is finally calculated to be 1.8. Similarly, the "A" of characteristic f2 yields an oversampling rate of 0.6, and then 1 is added so that the oversampling rate (r2) is finally calculated to be 1.6.

[0146] In the step (S130) of calculating sample weights based on an oversampling rate for data points, the oversampling rate may be used to calculate the number of times each data point must be sampled. Each data point x i For, sample weight n sample (x i ) can be calculated by multiplying the oversampling rates of the values ​​of the selected features. Sample weights for data points can be calculated as shown in the following mathematical formula.

[0147]

[0148] With reference to FIG. 3, a sample weight n is calculated by multiplying the oversampling rates r1 and r2 of features f1 and f2 at each data point in the input dataset. For example, a sample weight n=2.88 can be calculated by multiplying r11.8 and r21.6 of data point x1.

[0149] The sample weights calculated in this way are characteristic f i The number of times it is duplicated is determined, and for this purpose, the sample weights are rounded.

[0150] [Algorithm]

[0151] Step 3: Calculate Overall Sample Weight

[0152] for each data point x i do

[0153] Initialize n sample (x i )=1

[0154] for each feature in do

[0155]

[0156] end for

[0157]

[0158] end for

[0159] In the step (S140) of oversampling by duplicating data points in correspondence with the calculated sample weights, a data point to be duplicated is selected based on the magnitude of the sample weights and can be duplicated by the sample weight value. Referring to FIG. 3, among data points with the same characteristic f1, the data point having the highest sample weight can be selected as the target for duplication. For example, since data point 1 (1, A) has a sample weight of 2.88 and data point 2 (1, B) has a sample weight of 2.7, data point 1 can be selected as the target for duplication. In one embodiment of the present invention, the target for duplication is selected based on the characteristic f1 of each data point; however, alternatively, the magnitude of the sample weights among data points having the same characteristic based on characteristic f2 can be compared, and the data point having the largest value can be selected. Referring to FIG. 4, the sample weights of data points 1, 3, 4, and 10 selected as the target for duplication can be rounded to determine the number of times they will be duplicated. The sample weight of data point 1, 2.88, is rounded to 3, and accordingly, data point 1 can be duplicated to have a total of 3.

[0160] [Algorithm]

[0161] Step 4: Data Oversampling

[0162] for each data point x i do

[0163] for t=1 to do

[0164] Create duplicate

[0165] Calculate combined standard deviation:

[0166]

[0167] Sample noise ∈

[0168] Set

[0169] Add to the augmented dataset

[0170] end for

[0171] end for

[0172] Meanwhile, in another embodiment of the present invention, the step (S150) of adding noise to the duplicated data points may be further included.

[0173] To modify the replicated samples, i.e., the oversampled dataset, the target value y i Gaussian noise ε can be added. The noise is of a normal distribution. It can be extracted from, where the joint standard deviation (σ combined ) It can be calculated as (where k is a scaling factor that controls the magnitude of the noise). Through this step (S150), the new target value for each replicated sample is This becomes, and the generated data points It can be added to the augmented dataset. By adding noise to such an input dataset, overfitting of the existing model can be reduced and generalization performance can be improved.

[0174] The artificial intelligence system and oversampling method of the embodiments of the present invention as described above can effectively process categorical variables by preserving original feature values ​​during the oversampling process of a tabular input dataset, and can maintain data integrity without complex encoding or the synthesis and generation of categorical data.

[0175] To evaluate the artificial intelligence system and oversampling performance of the embodiments of the present invention, a tabular dataset applied to regression analysis may be used. The small dataset used to evaluate the performance of the embodiments includes ① an Analyzing Categorical dataset (AC), ② an Airfoil dataset (AF), ③ an Energy Efficiency dataset (EE), and ④ a Yacht Hydrodynamics dataset (YH).

[0176] The AC dataset contains 4,052 samples and 7 features. The AC dataset is a categorical analysis dataset used in the book *Analyzing Categorical Data*, and includes various datasets for scientific and educational purposes. The AF dataset, produced by NASA, contains 1,503 samples and 4 features related to airfoil performance, including frequency, angle of attack, and velocity. This dataset is generally used in studies predicting airfoil noise. The EE dataset is used in studies predicting energy efficiency based on building characteristics such as glass area and orientation, and it contains 768 samples and 8 features. The YH dataset contains 380 samples and 6 features related to yacht hydrodynamics, including hull shape and Froude number. This dataset is used in studies predicting residual drag per unit weight in various yacht designs.

[0177] The above four datasets were augmented by applying the oversampling method of the embodiments, and three deep learning models and two machine learning models were used to evaluate the augmented datasets. The deep learning models include MLP, ResNet, and FTTransformer, and the machine learning models may include CatBoost and XGBoost. In this case, the oversampling rate (r) was used as a sample weight in the machine learning models and as a gradient weight in the deep learning models. In the deep learning models, the oversampling rate may be included in the gradient of the loss function with respect to the model parameter θ. This is the slope of each data point at that ratio It can mean that it is scaled up to that extent.

[0178] With reference to FIG. 5, the hyperparameters of the deep learning model and the machine learning model were optimized using the original input dataset prior to oversampling to determine the baseline, and an evaluation was performed by inputting a dataset to which the oversampling method of the embodiments of the present invention was applied to the optimized model. In the evaluation, the scaling factor k was set to 0.01 and the number of important features d was set to 3, and 10 evaluations were conducted. As a result of calculating the average RMSE (Root mean square deviation) for the 10 training runs, the oversampled dataset according to the embodiments of the present invention showed a higher performance improvement compared to the general weighting method. In particular, a higher performance improvement was observed in small datasets with fewer than 1,000 samples, specifically in the EE and YH datasets. The oversampled EE dataset showed a very high performance improvement, reducing the RMSE of the XGBoost model by 43.02% from the baseline, and the oversampled YH dataset reduced the RMSE of the ResNet model by 52.46% from the baseline.

[0179] With reference to FIG. 6, further evaluation of embodiments of the present invention was performed. In one embodiment, the oversampling rate r is It is determined by (Proportionality). To evaluate the performance of the oversampling rate of this embodiment, a dataset was constructed by applying a modification to the oversampling rate, and a performance evaluation was conducted. The modification of the oversampling rate involved inverting and logarithmically inverting the oversampling rate. (Inversion: , logarithmic inversion: As a result of calculating the average RMSE (Root mean square deviation) for 10 training runs, it was confirmed that the oversampled result (Proportionality) according to the embodiments of the present invention showed a high performance improvement compared to the result of inverting and log-inverting the oversampling rate.

[0180] According to the artificial intelligence system and oversampling method of the embodiments of the present invention as described above, in a regression analysis task using a small tabular data set containing categorical input, significantly improved performance was derived from the input dataset improved by the oversampling method of the embodiments. These embodiments can provide practical alternatives in fields such as medicine, finance, and manufacturing, and better prediction accuracy can be derived by enhancing the learning process based on the embodiments.

[0181] In addition, the embodiments of the present invention can also be applied to the research and development of battery component materials, such as manganese-rich cathode materials. In the mass production project of manganese-rich cathode materials (Hero candidate project, acceleration of Mn-rich cathode material development), the applicant applied a data augmentation (oversampling) technique of the embodiments to compensate for the disadvantages of data biased by factor. By utilizing the oversampling method of the embodiments, two candidates were derived that showed improvement compared to the experimental reference sample in monocell (a basic battery unit consisting of a single layer of a positive and negative electrode) experiments among the complex doping conditions designed by AI. The feasibility and effectiveness of AI-based cathode material development based on the embodiments could be verified.

[0182] One embodiment of the present invention can be applied to an artificial intelligence model-based system for predicting new cathode materials. That is, one embodiment can be applied to a system that predicts new cathode materials by learning cathode material data containing doping elements using an artificial intelligence model, and can be applied to oversampling a data set.

[0183] Cathode materials are the components that form the cathode of secondary batteries and are a key material that influences the battery's energy density, output, lifespan, and stability. The main components of cathode materials include cathode active materials, which are composed of a combination of lithium (Li) and metallic components. The primary components of cathode active materials include nickel (Ni), cobalt (Co), manganese (Mn), and aluminum (Al). Among these cathode active materials, manganese-rich (Mn-rich) cathode materials—which reduce the content of expensive nickel and increase the proportion of manganese—are attracting attention as next-generation materials, leading to active research in this area. However, manganese-rich cathode materials suffer from low lithium-ion diffusion rates and conductivity, as well as the problem of manganese leaching during charging and discharging; to address these issues, coating and doping methods are being introduced. Coating is a method that enhances the lithium-ion velocity both inside and outside the active material, while doping is a method that enhances the velocity of lithium-ion and electrons within the active material. Doping elements for doping the cathode material may include Al (aluminum), Zr (zirconium), Mg (magnesium), Ti (titanium), B (boron), F (fluorine), W (tungsten), Mo (molybdenum), Ga (gallium), Vb (nitrogen group), Ca (calcium), etc.

[0184] Training data applied to the learning of the artificial intelligence model of one embodiment may include characteristic information of the cathode material. For example, the characteristic information of the cathode material may include analysis information including capacity, voltage, X-ray diffraction (XRD) and particle size distribution of the sintered product during charging and discharging, and battery performance information including capacity, lifespan, resistance increase, and gas generation. However, the characteristic information of the cathode material is not limited thereto and may include all information capable of distinguishing differences between materials.

[0185] An oversampling method of an artificial intelligence system according to one embodiment of the present invention can augment a dataset used for training an artificial intelligence model by oversampling it. An embodiment of the present invention can augment a cathode material dataset by oversampling it by calculating an oversampling rate for each data point of a cathode material, calculating a sample weight based on the oversampling rate for the data points, and oversampling the data points in correspondence with the calculated sample weight. In one embodiment, data points can be oversampled only for doping elements among the cathode materials. For example, if the cathode material dataset includes doping elements such as Co, Al, Nb, Mo, Ti, Zr, W, Mg, B, V, etc., the predictive performance of an artificial intelligence model trained based thereon can be improved by calculating an oversampling rate and a sample weight for each of these doping elements and oversampling the elemental characteristic information.

[0186] A computerized method for predicting a cathode material according to one embodiment of the present invention may include the step of inputting data-enhanced cathode material information into an artificial intelligence model executed by a processor for learning, and the step of predicting doping conditions of the cathode material using the learned artificial intelligence model executed by at least one processor.

[0187] Based on the cathode material and doping conditions derived according to the artificial intelligence model-based cathode material prediction system of one embodiment, an evolutionary algorithm-based cathode material can be designed based on capacity and lifespan. As a result of conducting coin cell (coin-shaped battery for evaluating cathode materials) experiments on candidate cathode materials designed according to one embodiment, samples exhibiting high performance in capacity, charge / discharge rate, and lifespan can be derived. When the sample serving as the standard for the experiment was set to tungsten (W), samples of tungsten / zirconium (W / Zr), tungsten / aluminum (W / Al), and tungsten / titanium (W / Ti) were derived.

[0188] Through monocell (a basic battery unit consisting of a single layer of a positive and negative electrode) experiments on three samples derived from coin cell experiments, two composite doping samples, specifically tungsten / zirconium (W / Zr) and tungsten / aluminum (W / Al), showed superiority over the reference sample in high temperature (45°C) life, as shown in [Table 1] below.

[0189] Category 0 cycle 100 cycle 200 cycle Lifetime (%) Capacity (mAh / g) Lifetime (%) Capacity (mAh / g) Lifetime (%) Capacity (mAh / g) Tungsten (W) Ref. 100.03 1.39 0.32 8.28 6.02 6.9 Tungsten / Zirconium (W / Zr) 100.09.92.37.69.56.7 Tungsten / Aluminum (W / Al) 100.00.01.57.58.96.7 Tungsten / Titanium (W / Ti) 100.01.38.67.74.96.6

[0190] Tungsten / zirconium (W / Zr) showed a lifespan of 92.3% in a 100-cycle monocell test and 89.5% in a 200-cycle test, confirming that its performance was improved compared to the reference. Additionally, tungsten / aluminum (W / Al) showed a lifespan of 91.5% in a 100-cycle monocell test and 88.9% in a 200-cycle test, confirming that its performance was improved compared to the reference. On the other hand, among the samples, tungsten / titanium (W / Ti) was found to have inferior performance compared to the reference in the monocell test.

[0191] In one embodiment of the present disclosure, for the development of a novel cathode material, a dataset containing cathode materials and doping elements is oversampled to form a dataset for training, and cathode material composite doping samples can be derived through an artificial intelligence model trained based on this, and the performance of the derived samples can be verified through coin cell and monocell experiments. That is, by using the novel cathode material prediction system and method of one embodiment of the present disclosure, results with high prediction reliability can be provided even for novel manganese-rich cathode materials.

[0192] One embodiment of the present invention can be applied to a protein function prediction system capable of predicting the function of a protein for proteins that do not have an EC number in the training set. That is, one embodiment can be applied to a system that predicts an EC number from the amino acid sequence of an enzyme using an artificial intelligence model and can be applied to oversampling a dataset.

[0193] Proteins are defined by sequences of one or more amino acids, and amino acids are organic compounds containing amino and carboxyl functional groups, as well as side chains (i.e., groups of atoms) specific to the amino acid. Protein folding refers to the physical process in which a sequence of amino acids folds into a three-dimensional configuration. The structure of a protein defines the three-dimensional arrangement of atoms within the protein's amino acid sequence after the protein has undergone protein embedding. When the sequence is connected by peptide bonds, amino acids may be referred to as amino acid residues.

[0194] Enzymes are biological catalysts that increase the rate of metabolism by binding to a substrate to form an enzyme-substrate complex, thereby lowering the activation energy of a chemical reaction. In some cases, they also perform a biological protective function by regulating the rate. Enzymes convert substrates into other molecules known as products. Like other catalysts, enzymes increase the reaction rate by lowering the activation energy of a chemical reaction. Some enzymes can accelerate the conversion of substrates into products by millions of times.

[0195] Enzymes are known to catalyze thousands of types of biochemical reactions. While most enzymes are proteins, some are RNA molecules with catalytic functions. RNA with catalytic functions is called ribozyme. The specificity of enzymes stems from their unique three-dimensional structure.

[0196] Enzymes are classified according to the numbering system of the Enzyme Commission (EC). The EC numbering system is a hierarchical classification system that classifies enzymes based on the reactions they catalyze, consisting of four levels that describe the enzyme's function. It is used to predict the EC number of an enzyme protein to identify and classify its catalytic activity, and this can be considered an important task that serves as a foundation for new drug development and various other protein or enzyme-related projects.

[0197] The EC (Enzyme Commission) number hierarchy is a numerical classification system designed to categorize enzyme proteins based on the chemical reactions they catalyze. EC numbers designate specific enzyme-catalyzed reactions, and different enzymes that catalyze the same reaction are assigned the same EC number. Through convergent evolution, the same EC number is assigned even when completely different protein foldings catalyze the same reaction. Therefore, the function of a protein enzyme can be determined using its EC number.

[0198] All EC numbers consist of the EC character and four digits separated by a period following the character. The first digit following the EC character is called the 1st level, followed by the 2nd through 4th levels. The four levels form a hierarchical structure, with the 1st level being the highest level and the 4th level being the most subdivided lower level.

[0199] The training data applied to the learning of the artificial intelligence model of one embodiment may include characteristic information of enzymes. For example, the characteristic information of enzymes includes the amino acid sequence, function, structure, post-translational regulation, and EC number of the enzyme, and such characteristic information may be obtained from databases provided by the NCBI (National Center for Biotechnology Information) maintained by the National Institutes of Health, the UniProt Knowledge Base (UniProtKB) provided by the Swiss Bioinformatics Institute, and the Swiss-Prot database. However, the characteristic information of enzymes is not limited to these and may include all information capable of distinguishing differences between enzymes. In addition, in some cases, multiple samples may be collected from enzymes associated with the input data to be processed by the trained model.

[0200] An oversampling method of an artificial intelligence system according to one embodiment of the present invention can augment a dataset used for training an artificial intelligence model by oversampling it from a plurality of collected samples. An embodiment of the present invention can augment a protein or enzyme dataset by oversampling it by calculating an oversampling rate for each of the data points including protein or enzyme characteristics, calculating sample weights based on the oversampling rates for the data points, and oversampling the data points in correspondence with the calculated sample weights. An artificial intelligence model can be trained based on the dataset thus augmented.

[0201] A computerized method for predicting protein function according to one embodiment of the present invention may include the steps of inputting data-enhanced protein amino acid sequence information into an artificial intelligence model executed by a processor, and predicting the EC number of a protein having said protein amino acid sequence information using an artificial intelligence model learned by hierarchical contrast learning executed by at least one processor.

[0202] Here, hierarchical contrast learning is trained using a training dataset containing training data that lacks information for one or more EC number hierarchy levels, and means that the trained model predicts an EC number at a level one step higher than the EC number hierarchy level without information. In this case, hierarchical contrast learning can be performed by converting the feature vector of the amino acid sequence of a protein into a representation vector using a multilayer perceptron with three hidden layers.

[0203] Hierarchical contrast learning of one embodiment may be performed by (i) using all dimensions of the representation when calculating the loss at the lowest level and using some dimensions of the representation used at lower levels when calculating the loss as one moves to higher levels, (ii) using only some dimensions of the representation when calculating the loss for each level, wherein some of the dimensions used for each level are used for calculation of the loss for other levels as well, (iii) dividing and using the dimensions used for each level so that they do not overlap, (iv) setting a number of learnable coefficient vectors equal in size to the representation vector equal to the number of levels and using the value obtained by multiplying the representation and the coefficient vector of each level when calculating the loss, (v) performing the task by placing as many multilayer perceptrons as there are levels, or through a combination of the above.

[0204] One or more artificial intelligence models included in the system of the present disclosure can infer the EC number of a query protein from a query protein. The inference is performed based on the distance between the query protein and the EC number.

[0205] Inference may include the steps of obtaining an embedding of each EC number from the EC numbers of a training dataset, using the Euclidean distance between the EC number embedding and the query protein embedding for inference, assuming that if an enzyme belongs to a specific progeny EC number, it belongs to the corresponding parent EC number, using the shortest distance to the progeny EC number as the distance to the parent EC number, and applying a maximum separation method to binarize labels according to distance.

[0206] In one embodiment of the present disclosure, a system and method for predicting protein function can be used to predict the function of a completely new enzyme protein for which an EC number has not previously been defined. In the absence of information regarding the EC number, it is impossible to predict the EC number of the lowest level, Level 4, but it is possible to predict the EC number of Level 3, which is one level higher.

[0207] In one embodiment of the present disclosure, hierarchical contrast learning can be performed using a training data set containing training data without information for one or more EC number hierarchy levels, and the trained model can predict an EC number at a level one step higher than the EC number hierarchy level without information with high confidence.

[0208] According to one embodiment of the present disclosure, higher levels of EC numbers can be predicted with improved performance even for protein enzymes that do not have EC numbers visible in the training data set. That is, by using the protein function prediction system and method of one embodiment of the present disclosure, results with high prediction reliability can be provided even for new proteins.

[0209] A computing system according to one embodiment of the present invention includes a user computing device and a server computing system, and each device and system can be connected to communicate through a communication unit.

[0210] A user computing device of one embodiment can perform the process of training an artificial intelligence model, retraining a model, or predicting or inferring the characteristics of a target (protein characteristics, material characteristics) by using an artificial intelligence model embedded in the device or an artificial intelligence model provided by a server computing system. In addition, the server computing device can provide a service for predicting target characteristics to the user computing device via an application or the web in response to a user's request through the user computing device.

[0211] A user computing device may include wearable devices, such as smartphones and tablet PCs, and all other types of computing devices, such as desktop computers. Such a user computing device includes at least one processor and memory. The memory may include one or more non-transient / transient computer-readable storage media and combinations thereof, and may include web storage of a server that performs the storage function of memory over the Internet. Such memory may store data and instructions necessary for at least one processor to perform operations of an application for learning / retraining an artificial intelligence model or performing target feature prediction.

[0212] Additionally, the user computing device may store at least one artificial intelligence model. For example, the user computing device may store machine learning models such as multiple neural networks, and other types of machine learning models including linear / non-linear models. Furthermore, the user computing device may store a prompt template as an input means to be used in the process of retraining the artificial intelligence model or predicting target features. That is, in one embodiment, the user computing device may perform target prediction based on received data by requesting execution through a prompt during the process of retraining or fine-tuning the model or predicting target features. Additionally, regarding the task requested through the user computing device, the server computing system may perform a process corresponding to the requested task through an embedded artificial intelligence model, and the execution result may be transmitted to the user computing device. Such a user computing device may include at least one user input unit that detects user input. For example, the user input unit may include a touch screen that detects a touch of a user's input medium (e.g., a finger or a stylus), an image sensor that detects the user's motion input, a microphone that detects the user's voice input, a button, a mouse, and a keyboard.

[0213] A server computing system includes at least one processor and memory, and at least one processor may be composed of at least one of a central processing unit (CPU), a graphics processing unit (GPU), ASICs (application specific integrated circuits), DSPs (digital signal processors), DSPDs (digital signal processing devices), PLDs (programmable logic devices), FPGAs (field programmable gate arrays), controllers, microcontrollers, microprocessors, and / or electrical units for performing other functions, or a plurality of electrically connected processors.

[0214] Memory may include one or more non-transient / transient computer-readable storage media and combinations thereof. Such memory may store artificial intelligence models, data, and instructions for at least one processor to train / retrain artificial intelligence models or perform target feature prediction. Memory may store neural networks or linear / non-linear models and may include feed-forward neural networks, deep neural networks, recurrent neural networks, and convolutional neural networks, etc. In one embodiment, the server computing system may further include a database which is a storage for continuously storing and managing raw data that forms the basis for training the artificial intelligence model, augmented data (oversampling data) for improving the prediction performance of the artificial intelligence model, etc. Such a database may include various forms of storage, including file systems and cloud storage. For example, a database may include at least one of a relational database that uses a structured query language (SQL) to define and manipulate data, a NoSQL database designed for flexibility and scalability to process unstructured and semi-structured data, a data warehouse optimized for querying and analysis by centralizing large volumes of data from multiple sources as a system used for reporting and data analysis, a data warehouse that stores large volumes of raw data in basic formats such as structured data, semi-structured data, and unstructured data, and a local storage device or a Network Attached Storage (NAS) that stores data in files in a format generally accessible by a computer operating system.

[0215] The communication unit performs wireless or wired communication between a user computing device and a server computing system, and the communication unit may use wireless communication systems according to methods such as eMBB, URLLC, MMTC, LTE, LTE-A, NR, UMTS, GSM, CDMA, WCDMA, TDMA, FDMA, OFDMA, SCFDMA, WiBro, WiFi, Bluetooth, NFC, GPS, or GNSS. In addition, the communication unit may use various wired communication systems such as USB, HDMI, RS-232 (Recommended Standard-232), POTS (Plain Old Telephone Service), Public Switched Telephone Network (PSTN), xDSL (x Digital Subscriber Line), RADSL (Rate Adaptive DSL), MDSL (Multi Rate DSL), VDSL (Very High Speed ​​DSL), UADSL (Universal Asymmetric DSL), HDSL (High Bit Rate DSL), and Local Area Network (LAN). In one embodiment, the communication unit may be configured regardless of the communication mode, such as wired or wireless, and may be configured with various communication networks, such as a Personal Area Network (PAN) or a Wide Area Network (WAN). Furthermore, the communication network may be the known World Wide Web (WWW), and may utilize wireless transmission technologies used for short-range communication, such as Infrared Data Association (IrDA) or Bluetooth.

[0216] A computing system of one embodiment can perform material property prediction or protein information prediction by utilizing an artificial intelligence model trained on a server computing system through a user computing device. In addition, a computing system of one embodiment can augment and strengthen the data set of the artificial intelligence model through a user computing device or through a server computing system to improve the performance of the trained artificial intelligence model.

[0217] One embodiment of the present invention can calculate an oversampling rate for each data point of a data set and calculate a sample weight based on the oversampling rate for the data points, and transmit a command to the server computing system to oversample the data points in correspondence with the calculated sample weight, so as to oversample a data set of a server computing system through a user computing device. Through such a command of the user computing device, an augmented oversampling data set is generated from an existing data set of the server computing system, and as the generated oversampling data set is included in the existing data set, the performance of predicting protein characteristics or cathode materials or cathode doping conditions can be further improved.

[0218] In addition, one embodiment of the present invention may improve the performance of an artificial intelligence model of a server computing system by augmenting a data set applied to an artificial intelligence model of a server computing system using an artificial intelligence model of a user computing device, and then transmitting the augmented oversampling data set or a training data set including the augmented oversampling data set to the server computing system.

[0219] A computerized method for predicting a cathode material according to one embodiment of the present invention may include the step of inputting data-enhanced cathode material information into an artificial intelligence model executed by a processor for learning, and the step of predicting doping conditions of the cathode material using the learned artificial intelligence model executed by at least one processor.

[0220] Meanwhile, embodiments of the present invention may be implemented as application-specific integrated circuits (ASICs) designed to suit specific application fields and special functions of devices.

[0221] Custom integrated circuits are also referred to as custom semiconductors. Unlike standard semiconductors, which have fixed specifications and can be applied to any electronic product or application as long as certain requirements are met, custom semiconductors are used for specific products or functions and are integrated circuits manufactured by semiconductor companies to meet specific orders. In other words, custom semiconductors are designed and manufactured to perform only the functions necessary for a specific device or feature. Custom semiconductors are broadly classified according to their design method into Full Custom ICs, which design and manufacture circuits from scratch to meet user requirements, and Semi-Custom ICs, which design and manufacture circuits using parts of a standardized design.

[0222] Application-specific semiconductors are primarily used in communication systems, high-performance computing systems, consumer electronics, automobiles, industrial automation, medical devices, the military, and the aerospace industry; recently, they are being applied to AI semiconductors that execute the large-scale computations required for AI implementation with high performance and power efficiency.

[0223] Application-specific semiconductors (ASICs) are used as core components in communication systems, such as network routers, switches, and modems, performing data packet processing, protocol conversion, and signal processing to provide high throughput and low latency. In high-performance computing systems, ASICs serve as key components for high-speed and parallel processing, while in consumer electronics—including digital cameras, smartphones, tablets, and game consoles—ASICs provide high-performance and low-power solutions required to perform specific functions. In the automotive industry, ASICs are used to control various electronic systems within vehicles, and in industrial automation systems, they provide solutions for high-precision control and high-performance processing.

[0224] An application-specific integrated circuit to which an embodiment of the present invention is applied includes a memory in which an individual memory interface (I / F) is implemented, and may include a plurality of function blocks that request memory access. Each function block may be a Direct Memory Access (DMA) function block, a processor, a video processor, a cache controller, a decompression block, or a data path block. The basic configuration of the application-specific integrated circuit may include a transistor that amplifies or switches an electrical signal, a logic gate which is a circuit that performs a logical function by combining transistors, a memory cell that stores data, an analog circuit which is a circuit that processes continuous voltage or current by combining transistors, and an Intellectual Property Core (IP Core) such as a microprocessor, DSP, or graphics core that is pre-designed to perform a specific function.

[0225] The ASIC may include an individual memory I / F that interfaces with individual memory and an embedded memory I / F that interfaces with embedded memory. The individual memory I / F is connected to each function block to receive memory access signals (e.g., control signals, address signals, and data signals) and, based on these input signals, can generate signals to control the individual memory. The embedded memory I / F is connected to each function block to receive memory access signals (e.g., control signals, address signals, and data signals) and, based on these input signals, can generate modified memory access signals to control the embedded memory. The individual memory I / F and the embedded memory I / F are designed within the memory control block of the ASIC to provide a memory control structure that can be flexibly applied to both the individual memory and the embedded memory.

[0226] Additionally, an application-specific integrated circuit (ASIC) for an artificial neural network (ANN) is composed of multiple neurons arranged in an array and multiple synapse circuits, each neuron being composed of a register, a microprocessor, and at least one input, and each synapse circuit being configured to include memory for storing synapse weights. Here, each neuron of the ASIC may be connected to at least one other neuron through one of the multiple synapse circuits.

[0227] Although the present disclosure has been described as generally being implementable by a computing device, a person skilled in the art will be well aware that the present disclosure may be implemented in combination with computer-executable instructions and / or other program modules that can be executed on one or more computers and / or as a combination of hardware and software.

[0228] Those skilled in the art of the present disclosure will understand that information and signals may be represented using any various different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols, and chips that may be referenced in the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof.

[0229] Those skilled in the art will understand that the various exemplary logic blocks, modules, processors, means, circuits, and algorithm steps described in connection with the embodiments disclosed herein may be implemented by electronic hardware, various forms of programs or design code (referred to herein as software for convenience), or a combination of all such. To clearly illustrate this interoperability between hardware and software, various exemplary components, blocks, modules, circuits, and steps have been generally described above in relation to their functions. Whether such functions are implemented as hardware or software depends on the design constraints imposed on the specific application and the overall system. Those skilled in the art may implement the functions described in various ways for each specific application, but such implementation decisions should not be interpreted as being outside the scope of this disclosure.

[0230] The various embodiments presented herein may be implemented as methods, devices, or articles manufactured using standard programming and / or engineering techniques. The term "article manufactured" includes a computer program, a carrier, or a medium accessible from any computer-readable storage device. For example, computer-readable storage media include, but are not limited to, magnetic storage devices (e.g., hard disks, floppy disks, magnetic strips, etc.), optical discs (e.g., CDs, DVDs, etc.), smart cards, and flash memory devices (e.g., EEPROMs, cards, sticks, key drives, etc.). Additionally, the various storage media presented herein include one or more devices and / or other machine-readable media for storing information.

[0231] It should be understood that the specific order or hierarchy of steps in the presented processes is an example of exemplary approaches. It should be understood that the specific order or hierarchy of steps in the processes may be rearranged within the scope of this disclosure based on design priorities. The appended method claims provide elements of various steps in a sample order, but do not imply being limited to the specific order or hierarchy presented.

[0232] Description of the presented embodiments is provided so that a person skilled in the art may use or practice the present disclosure. Various modifications to these embodiments will be apparent to a person skilled in the art, and the general principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Thus, the present disclosure is not limited to the embodiments presented herein, but should be interpreted in the broadest possible scope consistent with the principles and novel features presented herein.

Claims

1. As a system, At least one processor; and It includes at least one memory that stores instructions or information executed in at least one processor; and The operation performed by the instruction or information executed by the above at least one processor is, The method comprises an operation of calculating an oversampling rate for each of the data points of a dataset, an operation of calculating a sample weight based on the oversampling rate for the data points, and an operation of oversampling the data points in correspondence with the calculated sample weight. System.

2. In Claim 1, The above dataset includes multiple data points composed of feature vectors and target values ​​of feature vectors, System.

3. In Claim 1, The operation further includes calculating an importance score for the above data points, wherein the operation of calculating the importance score is, The operation of calculating an importance score by calculating the change amount of target values ​​of the above data points, and the operation of generating oversampling table data by selecting some of the data points having the highest importance score, System.

4. In Claim 1, The above oversampling rate is calculated as the ratio of the number of times each characteristic value of a plurality of characteristics included in the data points overlaps within the dataset, System.

5. In Claim 1, The above sample weight is calculated by multiplying multiple oversampling rates of the above data points, System.

6. In Claim 1, The above data points are duplicated by the above calculated sample weights and included in the dataset, System.

7. As a computerized method, A step of calculating the oversampling rate of each data point of a dataset stored in memory through a processor; A step of calculating sample weights based on the oversampling rate stored in the memory for each of the above data points through a processor; and A step comprising oversampling the data points in correspondence with the calculated sample weights stored in the memory through a processor, method.

8. In Claim 7, The above dataset is composed of multiple data points consisting of feature vectors and target values ​​of feature vectors, method.

9. In Claim 7, The method further includes the step of calculating importance scores for the above data points; and in the step of calculating the importance scores, An importance score is calculated by determining the change in target values ​​of the above data points, and some of the data points with the highest importance scores are selected to generate oversampling table data. method.

10. In Claim 7, The above oversampling rate is calculated as a ratio obtained by dividing the number of times each characteristic value of a plurality of characteristics included in the data points is duplicated within the dataset by the total number of data points. method.

11. In Claim 7, In the step of calculating the sample weights above, The above sample weight is calculated by multiplying multiple oversampling rates of the above data points and rounding the result, method.

12. In Claim 7, In the step of oversampling the above data points, The above data points are duplicated by the above calculated sample weights and included in the dataset, method.

13. An application-specific integrated circuit comprising a functional block including a memory in which information and instructions are stored and at least one processor that requests access to said memory, The above memory is, An instruction or information is stored that includes an operation of calculating an oversampling rate for each data point of a dataset, calculating a sample weight based on the oversampling rate for said data points, and oversampling said data points in correspondence with said sample weight. Custom Integrated Circuit.

14. As a computerized method, A step of calculating the oversampling rate of each data point of a dataset containing tabular cathode material characteristic information stored in memory through a processor; A step of calculating sample weights based on the oversampling rate stored in the memory for each of the above data points through a processor; and The method includes the step of oversampling the data points in correspondence with the calculated sample weights stored in the memory through the processor, At least some of the cathode material characteristic information of the above dataset is oversampled and stored in the above memory, method.

15. In Claim 14, The above dataset is composed of a plurality of data points consisting of a cathode material characteristic vector and a target value of the cathode material characteristic vector, and the above dataset includes cathode material characteristic information including a cathode active material and doping elements. method.

16. In Claim 15, The method further includes the step of calculating importance scores for the above data points; and in the step of calculating the importance scores, An importance score is calculated by determining the change in target values ​​of the above data points, and some of the data points with the highest importance scores are selected to generate oversampling table data. method.

17. In Claim 15, The above oversampling rate is calculated as the ratio obtained by dividing the number of times each characteristic value of the doping elements included in the data points is duplicated within the dataset by the total number of data points. method.

18. In Claim 14, In the step of calculating the sample weights above, The above sample weight is calculated by multiplying multiple oversampling rates of the above data points and rounding the result, method.

19. In Claim 15, In the step of oversampling the above data points, The above data points are duplicated by the above-calculated sample weights and included in the dataset, wherein some of the above-calculated doping elements are oversampled, method.

20. In Claim 14, The above cathode material characteristic information includes aluminum, zirconium, magnesium, titanium, boron, fluorine, tungsten, molybdenum, gallium, nitrogen group, and calcium as doping elements doping the cathode material, and calculates the oversampling rate and sample weight for each of the above doping elements so that the characteristic information of some of the above doping elements is oversampled. method.

21. In claim 20, A step of inputting data-enhanced cathode material characteristic information into an artificial intelligence model executed by a processor for learning; and The method further includes the step of predicting doping conditions of the cathode material using a learned artificial intelligence model executed by a processor; In the above prediction step, the predicted cathode material doping condition is derived as at least one of tungsten / zirconium (W / Zr), tungsten / aluminum (W / Al), or tungsten / titanium (W / Ti), method.

22. A computing system comprising a user computing device and a server computing system that performs a task corresponding to instructions or data received from the user computing device, The above server computing system comprises: a database storing a dataset for training an artificial intelligence model; a memory storing instructions and data for training the artificial intelligence model and performing the task; and at least one processor for performing the task according to the instructions or data in the memory. The above at least one processor is, Executing an operation of receiving an oversampling command for the dataset from the user computing device, calculating an oversampling rate for each of the data points of the dataset, calculating sample weights based on the oversampling rates for the data points, oversampling the data points in correspondence with the calculated sample weights, and providing the oversampling result to the user computing device. Computing system.

23. In Claim 22, The above at least one processor is, Further performing the operation of inputting an oversampled dataset into an artificial intelligence model executed by a processor for training, executing a prediction task based on instructions or data received from the user computing device, and providing the prediction task result to the user computing device for output. Computing system.

24. In Claim 22, The above dataset includes multiple data points composed of feature vectors and target values ​​of feature vectors, Computing system.

25. In Claim 22, The above at least one processor is, Further performing an operation to calculate importance scores for the above data points, wherein the operation to calculate importance scores includes an operation to calculate importance scores by calculating the change amount of target values ​​of the above data points, and an operation to generate oversampling table data by selecting some of the data points having high importance scores. Computing system.

26. In Claim 22, The above oversampling rate is calculated as a ratio obtained by dividing the number of times each characteristic value of a plurality of characteristics included in the data points is duplicated within the dataset by the total number of data points. Computing system.

27. In Claim 22, The above sample weight is calculated by multiplying multiple oversampling rates of the above data points and rounding the result, Computing system.

28. In Claim 22, The above data points are duplicated by the above calculated sample weights and included in the dataset, Computing system.

Images