AI RESEARCH
CliffSearch: Structured Agentic Co-Evolution over Theory and Code for Scientific Algorithm Discovery
arXiv CS.AI
•
ArXi:2604.01210v1 Announce Type: cross Scientific algorithm discovery is iterative: hypotheses are proposed, implemented, stress-tested, and revised. Current LLM-guided search systems accelerate proposal generation, but often under-represent scientific structure by optimizing code-only artifacts with weak correctness/originality gating.