framework
active
framework:evolver-wu-et-al-2025Evolver (Wu et al. 2025)
Self-evolving LLM agent system connecting offline strategy distillation with online retrieval
Neighborhood — ranked by edge-count
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- The update procedure (often an LLM) that converts agent execution evidence into harness updates
- Darwinian process of variation and selection; parallels with learning are central to the paper's argument.
- Machine learning approach using evolutionary processes to generate and select designs, used to blur the designed vs. evolved distinction
- Levin's model of how evolution repurposed ancient bioelectric machinery originally used for morphogenesis into neural control of behavior in 3D space.
- Fixed iterative protocol alternating between task-solving batches and harness evolution steps used across all experiments
- Evolutionary units change over evolutionary time and new units arise at new levels of organisation.claim0.729Emphasizes the dynamic nature of evolutionary individuality.
- Underlying theme: memory as communication between temporal and spatial agents.
- Field studying the relationship between evolution and development.