Deepbits Awarded NSF SBIR Phase I Grant for Pioneering AI-Powered Software Supply Chain Security Solution

September 2024 — Riverside, California — Deepbits has been awarded a prestigious National Science Foundation (NSF) Small Business Innovation Research (SBIR) Phase I grant for its project, “ReleaseChecker: Lastline Software Supply Chain Security via GPU-accelerated Binary Diffing.”

The project aims to address the urgent and growing threat of software supply chain attacks by introducing advanced AI-powered code diffing technology—capabilities that remain unavailable in existing software supply chain security solutions. By leveraging deep learning and GPU acceleration, ReleaseChecker promises to dramatically improve the speed and accuracy of analyzing changes between software releases, offering a vital “last line” of defense before deployment.

This innovation delivers several significant benefits:

Reduced Cybersecurity Costs: By automating and enhancing code analysis, ReleaseChecker allows U.S. businesses to allocate resources more efficiently, strengthening their global competitiveness.

Enhanced Software Supply Chain Security: The project is poised to significantly reduce the risk of cyberattacks, protecting sensitive data across government agencies, enterprises, critical infrastructure, and individual users.

Advancement of AI for Program Analysis: The project will deepen understanding of AI’s applications in cybersecurity—including binary code disassembling, function feature extraction, embedding, model training, and optimization. The new analysis pipeline established through this work will have broad applications across the cybersecurity industry.

Unlike traditional solutions that monitor each stage of the software supply chain, ReleaseChecker’s approach focuses on AI-powered binary code diffing—precisely identifying differences between software releases and leveraging software composition analysis and large language models (LLMs) to assess associated risks. This enables security teams to perform an effective final check before software deployment, substantially improving risk detection and compliance.

For more information about Deepbits and the ReleaseChecker project, please visit deepbits.com or contact info@deepbits.com.