Natural Language Embedded Programs for Hybrid Language Symbolic Reasoning

arXiv:2309.10814v2 Announce Type: replace
Abstract: How can we perform computations over natural language representations to solve tasks that require symbolic and numeric reasoning? We propose natural language embedded programs (NLEP) as a unifying framework for addressing math/symbolic reasoning, natural language understanding, and instruction following tasks. Our approach prompts a language model to generate full Python programs that define functions over data structures which contain natural language representations of structured knowledge. A Python interpreter then executes the generated code and prints …

abstract arxiv cs.cl embedded framework hybrid language language model language understanding math natural natural language prompts reasoning solve symbolic reasoning tasks type understanding