Optimized Routing in LEO Mega-Constellations Using AI
Discover how AI-driven intent compilation enhances constrained routing efficiency and safety in LEO mega-constellations with 99.8% delivery success.
Safe, Accountable GenAI for Women Denied Formal Education
Explore how safe, accountable GenAI empowers women banned from formal education with privacy, cultural sensitivity, and genuine learning support.
Automating Potential Discovery for the k-Server Conjecture
Explore k-server-bench, a breakthrough tool automating potential function discovery to advance the k-server conjecture in online algorithm research.
TraceSafe: Assessing LLM Guardrails in Multi-Step Tool Use
Explore TraceSafe-Bench, the first benchmark evaluating LLM guardrails' effectiveness in multi-step tool-calling trajectories for enhanced AI safety.
Advanced Mixture Proportion Estimation & Kernel Test Methods
Explore novel mixture proportion estimation and weakly-supervised kernel tests for conditional independence in AI and weakly supervised learning.