Learning Differentiable Programs with Admissible Neural Heuristics

Abstract

We study the problem of learning differentiable functions expressed as programs in a domain-specific language. Such programmatic models can offer benefits such as composability and interpretability; however, learning them requires optimizing over a combinatorial space of program "architectures". We frame this optimization problem as a search in a weighted graph whose paths encode top-down derivations of program syntax. Our key innovation is to view various classes of neural networks as continuous relaxations over the space of programs, which can then be used to complete any partial program. This relaxed program is differentiable and can be trained end-to-end, and the resulting training loss is an approximately admissible heuristic that can guide the combinatorial search. We instantiate our approach on top of the A-star algorithm and an iteratively deepened branch-and-bound search, and use these algorithms to learn programmatic classifiers in three sequence classification tasks. Our experiments show that the algorithms outperform state-of-the-art methods for program learning, and that they discover programmatic classifiers that yield natural interpretations and achieve competitive accuracy.

arXiv comment9 pages, published in NeurIPS 2020

References in corpus (12)

• Adam: A Method for Stochastic Optimization
• Neural Architecture Search with Reinforcement Learning
• DARTS: Differentiable Architecture Search
• DeepCoder: Learning to Write Programs
• Neural Programmer-Interpreters
• Synthesizing Programs for Images using Reinforced Adversarial Learning
• TerpreT: A Probabilistic Programming Language for Program Induction
• Neural Sketch Learning for Conditional Program Generation
• Neurally-Guided Procedural Models: Amortized Inference for Procedural Graphics Programs using Neural Networks
• Differentiable Programs with Neural Libraries
• HOUDINI: Lifelong Learning as Program Synthesis
• Write, Execute, Assess: Program Synthesis with a REPL

Cited by in corpus (1)

• Synthesizing Machine Learning Programs with PAC Guarantees via Statistical Sketching