WAMpy: Efficient Synthesis of Prolog Programs in Python

Dominik Magiera; Lukas Röhrig; Frank Jäkel

cs.PLOct 2, 2026

WAMpy: Efficient Synthesis of Prolog Programs in Python

Authors: Dominik Magiera, Lukas Röhrig, Frank Jäkel

Organizations: Centre for Cognitive Science & Institute of Psychology, Technische Universität Darmstadt, Germany

Abstract

We present WAMpy, a Python framework optimized for synthesizing Prolog programs. Unlike general-purpose Prolog systems, WAMpy targets workloads that repeatedly generate and evaluate small candidate programs. WAMpy compiles Prolog clauses into NumPy array-based WAM instructions and supports partial recompilation of hypotheses against fixed background knowledge. Performance-critical routines are accelerated using Numba just-in-time (JIT) compilation. In a benchmark of repeated compilation-and-evaluation workloads, WAMpy improves end-to-end performance compared with SWI-Prolog accessed from Python using Janus.

Figures & tables

Figure 1: WAMpy compilation and execution workflow. a. Pipeline from Prolog source code to query results. b. NumPy-based representation of a parsed term. c. Partial recompilation adds a candidate clause to precompiled background knowledge.

Figure 2: Performance benchmark of WAMpy and Janus SWI-Prolog. a. End-to-end runtime, including compilation and query execution. b. WAMpy compilation time for full-program compilation and partial hypothesis-only recompilation.

Program Synthesis

Program synthesis automatically constructs executable programs from examples, specifications, demonstrations, or natural-language descriptions.

See all 181 papers

Program-as-Weights: A Programming Paradigm for Fuzzy Functions
Jul 2, 2026Wentao Zhang, Liliana Hotsko, Woojeong Kim +3LLM-Based Program Synthesis Efficient Language Model Inference
University of Waterloo · 2Cornell University · 3Harvard University
Property-Guided LLM Program Synthesis for Planning
May 15, 2026André G. Pereira, Augusto B. Corrêa, Jendrik SeippSymbolic Planning Counterexample-Guided Refinement
Federal University of Rio Grande do Sul · Brazil · University of Oxford +3
ReaComp: Compiling LLM Reasoning into Symbolic Solvers for Efficient Program Synthesis
May 6, 2026Atharva Naik, Yash Mathur, Prakam +2Symbolic Reasoning LLM-Based Program Synthesis
Carnegie Mellon University · Independent Researcher

Jul 2, 2026J/K move · Enter open · S save

Program-as-Weights: A Programming Paradigm for Fuzzy Functions

Wentao Zhang, Liliana Hotsko, Woojeong Kim, Pengyu Nie+2

University of Waterloo · 2Cornell University · 3Harvard University

LLM-Based Program Synthesis Efficient Language Model Inference

Open paper AskPDF arXiv

Many everyday programming tasks resist clean rule-based implementation, such as alerting on important log lines, repairing malformed JSON, or ranking search results by intent, and are increasingly outsourced to large language model APIs at the cost of locality, reproducibility, and price. We propose fuzzy-function programming: compiling such a function from a natural-language specification into a compact, locally-executable neural artifact. We instantiate this paradigm with Program-as-Weights (PAW), in which a 4B compiler trained on FuzzyBench, a 10M-example dataset we release, emits parameter-efficient adapters for a frozen, lightweight interpreter. A 0.6B Qwen3 interpreter executing PAW programs matches the performance of direct prompting of Qwen3-32B, while using roughly one fiftieth of the inference memory and running at 30 tokens/s on a MacBook M3. PAW reframes the foundation model from a per-input problem solver into a tool builder: invoked once per function definition, it produces a small reusable artifact whose subsequent calls per function application are cheap and offline.

Abstract

Program Synthesis

Explore similar work

Program-as-Weights: A Programming Paradigm for Fuzzy Functions

Property-Guided LLM Program Synthesis for Planning

ReaComp: Compiling LLM Reasoning into Symbolic Solvers for Efficient Program Synthesis