This paper presents DroidCCT, a distributed framework for large-scale cryptographic compliance testing in the Android ecosystem, analyzing trillions of cryptographic operation samples from over half a billion devices to identify implementation weaknesses and bugs.
This talk presents FACADE, a novel self-supervised AI system used by Google to detect insider threats with high precision. FACADE uses a contrastive learning strategy trained solely on benign data, achieving a false positive rate below 0.01%.
This talk introduces the Sec-Gemini digital forensic agent that is able to automously perform timeline analysis and threat hunting with high accuracy on real-world compromised systems and direclty integrate with TimeSketch.
This talk announce the open-source release of the Phare Benchmark, an independent multi-lingual security and safety benchmark for large models alongside LMEval, a large model evaluation framework
We present GPAM the first side-channel attack model that generalizes across multiple cryptographic algorithms, implementations, and side-channel countermeasures without the need for manual tuning or trace preprocessing
This case-study explores the effectiveness of virtual reality (VR) for diversity, equity, and inclusion (DEI) training through the lens of a custom VR application developped to train Google employees."
How to build an accurate on-device RAG LLM system using Gemma, Ollama, USearch, and RETSim to answer questions about the characters of The Wizarding World of Harry Potter.
