Synthesizing Feature Extractors: An Agentic Approach for Algorithm Selection

Researchers at arXiv (2608.17170v1) introduced an automated approach using Large Language Models (LLMs) in an agentic check-fix-verify loop to synthesize executable Python feature extractors for constraint satisfaction problems, outperforming expert-curated features by up to 8.3 percentage points on the FLECC test set. The method, evaluated on vehicle routing, car sequencing, and fixed-length error-correcting codes with five solvers, consistently beat both mzn2feat and transformer-based trans2feat variants while remaining interpretable.

arXiv:2608.17170v1 Announce Type: new Abstract: Algorithm selection for constraint satisfaction problems requires extracting features that capture problem structure. Manually designing feature extractors demands deep domain expertise and quickly becomes a bottleneck when new problem classes appear. We present an automated approach that uses Large Language Models LLMs in an agentic check--fix--verify loop to synthesize executable Python scripts that act as interpretable, problem-specific feature extractors. Given a high-level MiniZinc model and an instance, the LLM agent generates code that constructs a typed graph representation and computes structural properties such as graph density, variable clustering, and constraint tightness. We evaluate our approach on three combinatorial problems vehicle routing, car sequencing, fixed-length error-correcting codes with a portfolio of five state-of-the-art solvers. The synthesized extractors yield algorithm selectors that consistently outperform both expert-curated mzn2feat features up to $8.3$ percentage points pp test-set accuracy on FLECC and the best transformer-based trans2feat variants. In the meanwhile, the synthesized feature extractors remain inspectable.