211 results about "Snippet" patented technology

A source code repair system generates a prompt including a few-shot examples of code changes made previously to correct a particular source code vulnerability. The prompt is given to a large language model to generate repair code for a source code snippet having the same vulnerability. Code changes made for a particular source code vulnerability, in the form of code diffs, are clustered into groups of closely related code change embeddings. A select few of the code diffs in each group having a closest mean of the cluster are used as the few-shot examples for the corresponding source code vulnerability.

Systems and methods for searching, summarizing, and automating workflows for financial related documents, and a user interface for displaying the same, are disclosed. The summaries are generated in response to a user query or through natural-language, conversational chat, and they provide the user with underlying supporting documentation to allow the user to follow-up and / or confirm as necessary. An AI-powered document search is performed to identify relevant documents and relevant snippets within the documents. The relevant snippets and / or documents are then summarized by one or more AI models. The models use metadata associated with the documents from the document search to rank the snippets for summarization. AI agents further automate the workflow.

A vulnerability detection and repair system utilize a classifier model to detect a software vulnerability in a source code snippet and the tokens in the source code snippet attributable to the vulnerability. A large language model is then given the vulnerable source code snippet, its vulnerability type, the vulnerability tokens, and a few-shot examples to determine whether or not the source code snippet includes the identified vulnerability. The few-shot examples include positive and negative samples of the type of vulnerability to guide the large language model towards the correct output.

Device, system, and method for automatically detecting and classifying personally identifiable information (PII) in documents and files. A method includes performing a deterministic rule-based search, in a plurality of stored documents, for PII data-items. If the deterministic rule-based search indicates that a particular document is more likely than not to contain a PII data-items then the method includes: extracting a textual snippet from the particular document, wherein the textual snippets surrounds the PII data-item; adding the textual snippet and the particular document to one or more training datasets utilized for training a Large Language Model (LLM) configured to find PII data-items in documents for Named Entity Recognition (NER) in those documents.

Techniques are described herein that are capable of providing a recommendation of AI-modified code in context of a developer tool. Based at least on code being developed in a developer tool, an interface element is provided in a user interface of the developer tool. The interface element is configured to receive a prompt that specifies a modification to be performed on the code. Based at least on receipt of the prompt, an AI model is automatically caused to perform the modification on at least a snippet of the code to provide a modified snippet. The modified snippet is processed using a language intelligence tool of the developer tool to provide a processed version of the modified snippet. A recommendation to replace the snippet in the code with the modified snippet is provided by causing the processed version of the modified snippet to be displayed via the user interface.

Embodiments described herein relate to systems and methods for automatically generating content for a generative answer interface of a collaboration platform. The system receives a natural language user input identifying corresponding blocks of text or snippets using a content extraction service. A prompt is generated using the blocks of text and is used to obtain a generative response. The generative response and links to corresponding content are displayed in the generative answer interface and can be inserted into content of the collaboration platform. The systems and methods described use a network architecture that includes a prompt generation service and a set of one or more purpose-configured large language model instances (LLMs) and / or other trained classifiers or natural language processors used to provide generative responses for content collaboration platforms.

A neural classifier model is used to detect cybersecurity vulnerabilities in the source code predicted by a deep learning code generation model having been trained on source code possibly containing security bugs. Upon the classifier model classifying a given source code snippet as likely containing a cybersecurity vulnerability, a proposed repair for the cybersecurity vulnerability is predicted from a neural decoder transformer model having been trained on non-vulnerable source code. The neural decoder transformer model is used to predict source code that repairs the cybersecurity vulnerability given the source code classified with a cybersecurity vulnerability.

The presently disclosed embodiments may include a computer readable medium including instructions that when executed by one or more processing devices cause the one or more processing devices to: receive an identification of at least one source audio or video with audio file, generate a textual transcript based on an audio component associated with the at least one source audio or video with audio file, edit the textual transcript to provide a formatted textual transcript, segment the formatted textual transcript into two or more segments, generate at least one summary snippet associated with the two or more segments, and cause the at least one summary snippet to be shown on a display together with a representation of the at least one source audio or video with audio file.

A retrieval-augmented neural transformer model with chunk cross-attention predicts a code review given a proposed source code change, represented as a code diff hunk, and a set of historical code review comments. The code diff hunk represents proposed edits to a source code snippet with its surrounding context that has not been changed. The historical code review comments are associated with code edits that are semantically similar to the proposed source code changes. The code diff hunk is partitioned into chunks which are used to find semantically similar historical code review comments. The set of historical code review comments is aggregated and used to guide the model in makings its predictions.

Methods and systems are provided for using generative artificial intelligence to evaluate fine-tuned language models. In embodiments described herein, natural language text snippets are generated via a generative language model based on corresponding data. A language model is fine-tuned into a fine-tuned language model via a language model fine-tuning component using the natural language text snippets and the corresponding data as training data. Independent natural language text snippets are generated via the generative language model based on the corresponding data. Each independent natural language text snippet is different than each corresponding natural language text snippet. An evaluation metric of the fine-tuned language model is generated via an evaluation component based on the independent natural language text snippets and the corresponding data.

The invention discloses an intelligent assistance method and system for software development, particularly relates to the field of software development assistance, is used for solving the problem of code snippet screening and recommendation, and aims to analyze an abstract syntax tree of code snippets through a systematic method, construct a dependency graph and a time sequence dependency graph, and calculate early warning isolation and a functional functional rate. Screening high-value code snippets by combining the indexes; converting the code snippets into high-dimensional vectors by using a deep learning model, capturing grammar and semantic features, understanding the intention of a developer in real time through a situation analysis model, and adjusting a code recommendation strategy; clustering and classifying the code snippets by using an unsupervised learning algorithm to form a code snippet library with similar semantics and related functions; the accuracy of classification, clustering and similarity calculation of the code snippets is improved, code search and recommendation are effectively supported, the development process is optimized, the reuse rate and the system stability are improved, and the maintenance cost is reduced.

The invention relates to artificial intelligence and network security technologies, in particular to a dynamic malicious software detection model black box confrontation disturbance method based on a large model. According to the method, a dynamic malicious software detection model is given, and an API sequence sample set is constructed; training by adopting a machine learning algorithm to obtain a proxy malicious software detection model; constructing a generator, obtaining a trained generator by utilizing an agent malicious software detection model through adversarial training, and performing feature space adversarial disturbance on the API feature vector of the malicious software sample by utilizing the trained generator to obtain an API feature vector after adversarial disturbance; obtaining an API difference vector based on the API feature vector of the malicious software sample and the API feature vector after confrontation disturbance; and generating a function code snippet by using the large model, and modifying the source code of the malicious software sample according to the function code snippet to obtain a final anti-disturbance code. According to the method, the work of manual coding is avoided, and the attack resisting efficiency is greatly improved.

The invention provides a code generation method which is executed by a code generation system, the code generation system is used for automatically generating a test code, and the method comprises the following steps: receiving a test step description input by a user, retrieving a text code pair vector library and a test method vector library according to the test step description, the text code pair vector library comprises description texts and vectors formed by corresponding test codes, and the test method vector library comprises vectors formed by the test method calling examples; and according to the test code snippets and the test methods corresponding to the similar test step descriptions, reasoning through a large language model (LLM), generating test codes corresponding to the test step descriptions, and presenting the test codes corresponding to the test step descriptions to a user. According to the method, retrieval and LLM reasoning are combined, the complex test code can be automatically generated, and the automatic development efficiency of the complex test scene is improved.

The present invention relates to a snippet information generation method and apparatus, a search result display method and apparatus, a device, and a medium. The search result display method includes: in response to search content, obtaining a first research result; displaying, in a search result page, snippet information corresponding to the first search result, wherein the snippet information is generated on the basis of detailed content and snippet attributes of the first search result, the snippet attributes comprise at least one of the snippet row number, the snippet area size, and snippet typesetting information, and the snippet attributes are determined on the basis of key content matched with the search content in the detailed content.

The invention discloses a code review method, system and device, a storage medium and a program product, and the method comprises the steps: cleaning a to-be-reviewed code snippet, and obtaining an incremental code snippet and position information; vectorizing the incremental code snippets; semantic query is conducted on the vectorized incremental code snippet data based on an ES database according to the cosine law, development rules of similar semantics are obtained and ranked from high to low according to similarity values, and a preset number of first several snippets are obtained; recombining cue words, wherein the recombining cue words comprise incremental code snippets and a preset number of development rules; and sending the recombined cue word to a general generative large model to obtain an examination suggestion. The code snippets to be examined are pushed to the gilab warehouse by a developer, and are cleaned by the gilab warehouse through webhook triggering. According to the review method, the whole process is automatic, the review efficiency is high, review can be conducted according to the development rules in the team, knowledge sharing is achieved, the review result is objective, and the problem that the text exceeds the upper limit is avoided.

Method for prioritizing code violations in a computer program, using machine learning includes: analyzing the computer program for code violations; extracting code snippets containing violations from the computer program; training a machine learning model to differentiate between vulnerable and non-vulnerable code in the extracted code snippets; inputting the extracted code snippets to a trained machine learning model to assign a vulnerability probability score to each snippet, wherein each vulnerability probability score indicates a severity of the violation for a respective snippet; ranking the code snippets based on their respective vulnerability probability score, wherein a higher score indicates a higher likelihood of causing severe vulnerabilities; and displaying the ranked code snippets to be fixed for their code violations.

A retrieval-augmented neural transformer model with chunk cross-attention predicts a code review given a proposed source code change, represented as a code diff hunk, and a set of historical code review comments. The code diff hunk represents proposed edits to a source code snippet with its surrounding context that has not been changed. The historical code review comments are associated with code edits that are semantically similar to the proposed source code changes. The code diff hunk is partitioned into chunks which are used to find semantically similar historical code review comments. The set of historical code review comments is aggregated and used to guide the model in makings its predictions.

Embodiments of the present disclosure provide a code generation method, a device, and a storage medium. In the embodiments of the present disclosure, a custom retrieval library can be constructed, wherein the custom retrieval library comprises sub-snippets segmented from reference code snippets; on this basis, a target sub-snippet meeting a preset similarity requirement can be retrieved for an uncompleted code from the custom retrieval library; and then a code context of the target sub-snippet can be acquired to generate a retrieval result. Since the custom retrieval library is at the sub-snippet level, the code diversity in the custom retrieval library can be effectively enhanced, and various code variants can be covered, so that the retrieval precision can be effectively improved, effectively improving the quality of retrieval results. Moreover, prompt information can be constructed on the basis of the generated retrieval result and input into a code generation model.

The invention is applicable to the technical field of code development, and provides a code modification method based on a CURD scene, which comprises the following steps: acquiring a business demand of code modification, and acquiring a data model required by code modification according to the business demand; based on the data model, obtaining a code file related to a CURD scene during code modification; analyzing a calling relationship among functions related to the CURD scene in the code file, and constructing a function calling graph; on the basis of the service requirements and the data model, service code snippets are generated, the service code snippets are combined and modified to the code file, code modification of the service requirements can be processed, and development efficiency and code quality are further improved.

The invention discloses a general algorithm reflection evolution (MHRE) method based on a large language model. The MHRE method comprises the steps of individual coding, population construction and population iterative optimization. In the individual coding stage, the large language model generates code snippets containing performance functions and architecture functions, and the snippets define basic structures and behaviors of population individuals. In the population construction stage, a prompt word of a specific task is used for further refining code snippets, and initial population individuals are formed. In the population iterative optimization stage, through a cross evolution module, a cooperative evolution module, a framework upgrading module and other modules, iterative optimization of population individuals is achieved in combination with a generation-standardization-evaluation-selection loop module (GSES), and algorithm individuals with the optimal performance are generated. The MHRE method can automatically generate, optimize and evaluate a plurality of algorithms, improves the adaptability, universality and optimization effect of the algorithms, and is especially suitable for complex optimization tasks such as industrial production, resource allocation and logistics path planning.

A vulnerability detection and repair system utilize a classifier model to detect a software vulnerability in a source code snippet and the tokens in the source code snippet attributable to the vulnerability. A large language model is then given the vulnerable source code snippet, its vulnerability type, the vulnerability tokens, and a few-shot examples to determine whether or not the source code snippet includes the identified vulnerability. The few-shot examples include positive and negative samples of the type of vulnerability to guide the large language model towards the correct output.

The invention provides a code searching method and device, electronic equipment and a storage medium, and relates to the technical field of retrieval. The method comprises the steps that in response to a received retrieval statement of a natural language, a natural language annotation corresponding to the retrieval statement is searched in a natural language annotation set of a code library, and the natural language annotation is used for indicating functions of functions in function code snippets in the code library; and if the natural language annotation is found, determining a function code snippet corresponding to the natural language annotation in the code library. According to the embodiment, codes do not need to be adopted as search terms, natural languages can be adopted for searching, searching personnel are more friendly, the searching habits of the personnel are better met, searching can be conducted through the functions, and the needed functions can be searched easily.

A source code repair system generates a prompt including a few-shot examples of code changes made previously to correct a particular source code vulnerability. The prompt is given to a large language model to generate repair code for a source code snippet having the same vulnerability. Code changes made for a particular source code vulnerability, in the form of code diffs, are clustered into groups of closely related code change embeddings. A select few of the code diffs in each group having a closest mean of the cluster are used as the few-shot examples for the corresponding source code vulnerability.

Implementations set forth herein relate to an automated assistant that can selectively communicate audio data to a recipient when a user solicits the automated assistant to send a text message to the recipient. The audio data can include a snippet of audio that characterizes content of the text message, and the automated assistant can communicate the audio data to the recipient when score data for a speech recognition hypothesis does not satisfy a confidence threshold. The score data can correspond to an entirety of content of a text message and / or speech recognition hypothesis, and / or less than an entirety of the content. A recipient device can optionally re-process the audio data using a model that is associated with the recipient device. This can provide more accurate transcripts in some instances, thereby improving accuracy of communications and decreasing a number of corrective messages sent between users.

In some implementations, a remediation system may receive, a tracking system, a set of tickets associated with at least one data record. The remediation system may search, in the at least one data record and using a set of contexts indicated in the set of tickets, for a set of sensitive data snippets. The remediation system may generate a set of proposed remediations corresponding to the set of sensitive data snippets. The remediation system ma6y output, for each sensitive data snippet in the set of sensitive data snippets, a corresponding proposed remediation in the set of proposed remediations. The remediation system may selectively apply the set of proposed remediations based on inputs from a user.

A vulnerability detection and repair system utilizes a classifier model to detect software vulnerabilities in source code fragments and lexical elements attributable to the vulnerabilities in the source code fragments. The vulnerable source code fragment, its vulnerability type, vulnerability lexical elements, and few sample examples are then given to a large language model to determine whether the source code fragment includes the identified vulnerability. The few sample examples include positive and negative samples of vulnerability types to direct the large language model towards correct outputs.

An assert statement generator employs a neural transformer model with attention to generate candidate assert statements for a unit test method that tests a focal method. The neural transformer model is pre-trained with source code programs and natural language text and fine-tuned with test-assert triplets. A test-assert triplet includes a source code snippet that includes: (1) a unit test method with an assert placeholder; (2) the focal method; and (3) a corresponding assert statement. In this manner, the neural transformer model is trained to learn the semantics and statistical properties of a natural language, the syntax of a programming language, and the relationships between the code elements of the programming language and the syntax of an assert statement.

Systems and methods are disclosed for automatically extracting relevant information from various source document that include tabular data. The tabular data in various forms can be received as mixed with other dissimilar data. Tabular data can appear in different orientations, document types, and can be fragmented horizontally or vertically. The proposed technique automatically detects table header data in certain regions of the received source document and associates values to the extracted headers. The proposed system is capable of combining different snippets of smaller tables into a single cohesive and monolithic table with headers designated by a set of keywords and all the values in the various columns (and / or rows) included under or along proper headers.

The invention provides an applet permission security detection method and system, and relates to software engineering.The method comprises the steps that API test code snippets of a super application are generated based on a large language model, the large language model is guided according to an API dependency tree to generate a plurality of API test code snippets, and executable test codes are constructed; dynamically executing the executable test code in the simulation environment, and obtaining the actual permission configuration of the applet according to the dynamic execution result; obtaining standard permission configuration of the super application, comparing the actual permission configuration with the standard permission configuration, and if the situation that the actual permission configuration does not accord with the standard permission configuration occurs, determining that the super application has an API permission exception problem; the method comprises the following steps: constructing an abstract syntax tree according to an applet source code, performing granularity optimization on the basis of the abstract syntax tree to construct an abstract method tree, searching an abnormal API (Application Program Interface) with a permission defect in the abstract method tree, determining a function body name where the abnormal API is located, performing front-end and rear-end dual confirmation according to the function body name, and determining a calling condition and a binding condition of the abnormal API. Accurate detection of the permission defect in the super application is realized.

One or more new coding instructions are generated using a language model (LM) prompted to perform one or more genetic operations on one or more seed coding instructions of an initial set of coding instruction-snippet pairs. One or more respective coding snippets are generated to implement the one or more new coding instructions using a LM prompted to generate coding snippets for the one or more new coding instructions. A generational set of coding instruction-snippet pairs comprising the initial set of coding instruction-snippet pairs and a new set of coding instruction-snippet pairs comprising the one or more new coding instructions and the one or more respective coding snippets is created.