Program-as-Weights: A Programming Paradigm for Fuzzy Functions
2026-07-02 • Machine Learning
Machine LearningArtificial IntelligenceComputation and Language
AI summaryⓘ
The authors introduce a new way to turn natural language instructions into small, efficient computer programs called fuzzy-function programming. They created a system named Program-as-Weights (PAW) that compiles these instructions into compact neural network adapters. This allows running complex tasks on a small device like a laptop much faster and using less memory than directly asking large language models every time. Their approach makes these models into tool builders that produce reusable programs instead of solving each problem from scratch.
fuzzy-function programminglarge language modelsneural adaptersparameter-efficient tuningProgram-as-Weights (PAW)natural language specificationmodel compilationoffline inferencefoundation modelsresource-efficient AI
Authors
Wentao Zhang, Liliana Hotsko, Woojeong Kim, Pengyu Nie, Stuart Shieber, Yuntian Deng
Abstract
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.