Adaptive Neural Compilation

Abstract

This paper proposes an adaptive neural-compilation framework to address the problem of efficient program learning. Traditional code optimisation strategies used in compilers are based on applying pre-specified set of transformations that make the code faster to execute without changing its semantics. In contrast, our work involves adapting programs to make them more efficient while considering correctness only on a target input distribution. Our approach is inspired by the recent works on differentiable representations of programs. We show that it is possible to compile programs written in a low-level language to a differentiable representation. We also show how programs in this representation can be optimised to make them efficient on a target distribution of inputs. Experimental results demonstrate that our approach enables learning specifically-tuned algorithms for given data distributions with a high success rate.

Submitted to NIPS 2016, code and supplementary materials will be available on author's page

References in corpus (4)

• Adding Gradient Noise Improves Learning for Very Deep Networks
• Neural GPUs Learn Algorithms
• Reinforcement Learning Neural Turing Machines - Revised
• Learning Simple Algorithms from Examples

Cited by in corpus (11)

• Learning to Perform Local Rewriting for Combinatorial Optimization
• Programming with a Differentiable Forth Interpreter
• Not all bytes are equal: Neural byte sieve for fuzzing
• TerpreT: A Probabilistic Programming Language for Program Induction
• Learn&Fuzz: Machine Learning for Input Fuzzing
• Learning Compositional Neural Programs with Recursive Tree Search and Planning
• Learning to superoptimize programs
• Program Language Translation Using a Grammar-Driven Tree-to-Tree Model
• Differentiable Functional Program Interpreters
• Program Synthesis Through Reinforcement Learning Guided Tree Search
• Summary - TerpreT: A Probabilistic Programming Language for Program Induction