Are Self-Evolving Agents Continual Learners?
Evaluating three representative methods spanning experience retrieval and self-modification (ReasoningBank, GEPA, and Meta-Harness), we find that although self-evolution improves aggregate performance on most streams, higher plasticity consistently comes at the cost of stability.
ProofPaper ↗
Key points
- Large language model agents are often deployed in environments where new knowledge, tools, and workflows keep arriving.
- Such an agent is therefore expected to adapt continually, learning new tasks as they arrive without forgetting earlier ones.
- Nonparametric self-evolving agents are a natural alternative: they adapt by rewriting their own harness, such as prompts, memory, and skills, from task feedback while the model weights stay frozen.
- We study whether such agents are effective continual learners across four drifting task streams that shift in topic, workflow, or domain (built from BIRD, BFCL, and ScienceWorld).
Sources (1)
- [1]Are Self-Evolving Agents Continual Learners?arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 19, 06:12 AM
Evaluating three representative methods spanning experience retrieval and self-modification (ReasoningBank, GEPA, and Meta-Harness), we find that although self-evolution improves aggregate performance on most streams, higher plasticity consistently comes at the cost of stability.
Large language model agents are often deployed in environments where new knowledge, tools, and workflows keep arriving.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 19, 2026Ropedia W0: Scaling Loco-Manipulation World Models across Embodiments with Human Experience
- Oct 19, 2026Agent World Model in the Wild: Computer-Use Agent Environment Synthesis
- Oct 19, 2026AnyGoal: On World Model Generalization Beyond Expert Rollouts
- Oct 19, 2026SpatialScout: Can VLM Agents Actively Explore and Map 3D Scenes?
- Oct 18, 2026Google Open-Sources Harness2, a Recursive Agent Harnessing Framework
- Oct 18, 2026CoEvo-RSI: When Harnesses Evolve, Which Models Should Follow?