adaptation

2 items

ARTICLEDEV.to AI·7d ago

The agent setup that updates itself. What self-evolving actually means in practice

The article argues that true self-evolving AI agents are practical and "boring": they notice changes, learn, and avoid relearning, unlike theatrical demos. It emphasizes that agent setups decay quickly and must continuously adapt to real-world changes to remain effective.

software development adaptation self-evolution AI agents

RESEARCHarXiv CS.AI·29d ago

CASCADE: Case-Based Continual Adaptation for Large Language Models During Deployment

This paper introduces Deployment-Time Learning (DTL) as a new stage for LLMs, allowing them to continually adapt from experience post-training without modifying core parameters. It presents CASCADE, a framework that uses an explicit, evolving episodic memory for LLM agents, formalizing experience reuse as a contextual bandit problem with no-regret guarantees.

LLMs adaptation machine learning AI deployment