Rigorous Explainability in AI via Feature Attribution
Explore rigorous symbolic methods for trustworthy feature attribution in AI, enhancing transparency and reliability over non-symbolic approaches.
Experience Compression Spectrum: Boosting LLM Agent Efficiency
Discover the Experience Compression Spectrum, unifying memory, skills, and rules to optimize LLM agents' efficiency and reduce computational overhead.
Discover And Prove: Advanced Hard Mode Theorem Proving
Explore Discover And Prove, an open-source framework enhancing automated theorem proving in Lean 4 with Hard Mode challenges and LLM integration.
Stein Variational Method for Combinatorial Optimization
Discover how Stein variational gradient descent improves exploration in combinatorial black-box optimization for large-scale, complex problems.
KWBench: Benchmarking Unprompted Problem Recognition in AI
Discover KWBench, the first benchmark to measure unprompted problem recognition in large language models across diverse professional tasks.