Why some agents fail to improve: reuse rates alone don't explain learning gains
anirudhg9119 · x · 2026-10-09
Core conclusions from the same Agent Plasticity thread: low reuse of learned artifacts correlates with weak gains, but even agents reusing artifacts in 94–98% of relevant decisions improve by very different amounts, and for strong improvers 83–99% of remaining failures happen while an artifact is in use. The authors propose Agent Plasticity—held-out gain per unit of learning cost—showing capability ≠ learning efficiency. See the arXiv paper entry in this thread for full details.
More from Research
- RWTH Aachen's ARROW unifies 3D reconstruction and point tracking from any RGB inputs, sets new SOTA — kwangmoo_yi · 2026-10-09
- HAIPS 2026 workshop lands at COLM tomorrow with top AI privacy researchers — tianshi_li · 2026-10-09
- Tighter Bayes error bounds via generalized Bhattacharyya and Chernoff means — FrnkNlsn · 2026-10-09
- PAMI: part-anchored motion lifts contact recall 14.5% in text-to-HOI generation — _akhaliq · 2026-10-09
- Study finds LLM novelty judges are unstable: scores swing wildly with evaluation design choices — _akhaliq · 2026-10-09
- Google's AMIE lands in The Lancet: 90% match with doctors' final diagnoses in real clinics — sundarpichai · 2026-10-09