Abstract: A computer-implemented method for detecting malicious code across different programming languages receives input source code, generates semantic embeddings for the input code and cross-language generated code, compares the embeddings to produce a comparison result, and analyzes the result to identify potential malicious code. The method leverages semantic analysis techniques and language-specific embedding models to enable efficient and accurate detection of security threats across language boundaries. By focusing on the semantic essence of the code rather than its literal text representation, the method overcomes limitations of traditional malicious code detection approaches, providing a robust, scalable, and secure solution for managing code security in multi-language software development environments.

Abstract: A data management system, also referred to as a “smart data room,” streamlines the process of requesting, collecting, and analyzing data through a structured Data Request List (DRL). The system facilitates interactions between requestors and respondents via an automated platform that supports the creation, submission, and review of data requests and responses. Key features include the ability to save and reuse DRLs and responses as templates, real-time notifications for tracking response progress, secure data deletion upon request, and advanced data validation using machine learning techniques to ensure the relevance and adequacy of the information collected. Additionally, the system offers capabilities for conducting live interviews using a real-time avatar, with responses being recorded and analyzed for relevance and adequacy. The system enhances data integrity and efficiency in data collection processes, making it particularly useful in environments where accurate and timely data collection is critical.

Abstract: A data management system, also referred to as a “smart data room,” streamlines the process of requesting, collecting, and analyzing data through a structured Data Request List (DRL). The system facilitates interactions between requestors and respondents via an automated platform that supports the creation, submission, and review of data requests and responses. Key features include the ability to save and reuse DRLs and responses as templates, real-time notifications for tracking response progress, secure data deletion upon request, and advanced data validation using machine learning techniques to ensure the relevance and adequacy of the information collected. Additionally, the system offers capabilities for conducting live interviews using a real-time avatar, with responses being recorded and analyzed for relevance and adequacy. The system enhances data integrity and efficiency in data collection processes, making it particularly useful in environments where accurate and timely data collection is critical.

Abstract: One embodiment of the present invention relates to a computer-automated system and method for evaluating software development metrics to enhance the diligence process and detect source code plagiarism. The system integrates technical and financial metrics to assess the productivity and economic impact of software developers. It includes a plurality of data sources, such as work product data sources containing source code and financial data sources detailing compensation. The system processes and analyzes this data to generate outputs reflecting worker performance and financial efficiency. Key features include complexity analysis of source code, sentiment analysis, and outlier detection in financial transactions. The system provides synthesized outputs, such as dashboards and reports, which are reviewed and approved before being shared with requesters.

Abstract: One embodiment of the present invention relates to a computer-automated system and method for identifying copyrighted source code embedded within other source code files, utilizing advanced semantic analysis techniques. This embodiment of the invention addresses the challenge of detecting both literal and non-literal copies of copyrighted code, including instances where the code has been modified in non-semantic ways, such as through renaming variables, changing formatting, or rearranging code blocks. This embodiment creates semantic embeddings of source code using a large language model (LLM). Each segment of source code is transformed into a high-dimensional vector that captures its semantic essence, rather than its literal text. These vectors are then compared using sophisticated similarity metrics, such as cosine similarity or L2 distance, to determine the likelihood of copyright infringement.