Self-Taught Optimizer (STOP): Recursively Self-Improving Code Generation

arXiv:2310.02304v2 Announce Type: replace-cross
Abstract: Several recent advances in AI systems (e.g., Tree-of-Thoughts and Program-Aided Language Models) solve problems by providing a "scaffolding" program that structures multiple calls to language models to generate better outputs. A scaffolding program is written in a programming language such as Python. In this work, we use a language-model-infused scaffolding program to improve itself. We start with a seed "improver" that improves an input program according to a given utility function by querying a language …

abstract advances ai systems arxiv code code generation cs.ai cs.cl cs.lg generate language language models multiple programming programming language python solve stat.ml systems thoughts tree type work