Research

APEX-EM: Non-Parametric Online Learning for Autonomous Agents via Structured Procedural-Episodic Experience Replay

APEX-EM gives Claude agents persistent procedural memory to reuse solutions for structurally similar tasks without retraining, achieving 89.6% on code generation benchmarks (+48 points over baselines).

Monday, April 6, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.CL (Computation & Language)BY sys://pipeline

APEX-EM is a non-parametric online learning framework that gives LLM-based autonomous agents persistent procedural memory, allowing them to reuse solutions for structurally similar tasks without modifying model weights. Evaluated on Claude Sonnet 4.5 and Opus 4.5, it achieves massive gains: 89.6% on KGQAGen-10k (+48.3pp), 83.3% on BigCodeBench (+29.4pp), exceeding oracle baselines. This directly improves agent reliability and efficiency for code generation and autonomous reasoning.

Read original at arXiv CS.CL (Computation & Language)

Nation-states want to cause harm, not just steal cash - stop handing your cyber defenses to the cheapest contractor

UK National Cyber Security Centre CEO warns that China now represents a peer-level competitor in cyberspace with sophisticated state-sponsored attacks, citing an average of four nationally significant incidents per we...