WAMpy: Efficient Synthesis of Prolog Programs in Python
WAMpy is built for the tight loop of generating and testing many small Prolog candidates from Python.
The framework compiles Prolog clauses into NumPy array-based WAM instructions and uses Numba JIT for performance-critical routines. It also supports partial recompilation when hypotheses change against fixed background knowledge. The paper reports faster end-to-end performance than SWI-Prolog called from Python through Janus on repeated compilation-and-evaluation benchmarks. Accepted as an IJCLR 2026 demo. ArXiv · AI/CL/LG's note
The framework compiles Prolog clauses into NumPy array-based WAM instructions and uses Numba JIT for performance-critical routines. It also supports partial recompilation when hypotheses change against fixed background knowledge. The paper reports faster end-to-end performance than SWI-Prolog called from Python through Janus on repeated compilation-and-evaluation benchmarks. Accepted as an IJCLR 2026 demo. ArXiv · AI/CL/LG's note
score 4