Data-Efficient Learning with Neural Programs

TL;DR

This paper introduces ISED, an algorithm for learning neural programs that efficiently estimates gradients of black-box components using only input-output data, enabling effective training of composite models involving LLMs and neurosymbolic systems.

Contribution

The paper presents ISED, a novel gradient estimation algorithm for neural programs that works with black-box components and improves data efficiency over existing methods.

Findings

ISED achieves comparable accuracy to state-of-the-art neurosymbolic frameworks.

ISED is more data- and sample-efficient than prior gradient approximation methods.

Evaluation includes benchmarks with GPT-4 and neurosymbolic tasks.

Abstract

Many computational tasks can be naturally expressed as a composition of a DNN followed by a program written in a traditional programming language or an API call to an LLM. We call such composites "neural programs" and focus on the problem of learning the DNN parameters when the training data consist of end-to-end input-output labels for the composite. When the program is written in a differentiable logic programming language, techniques from neurosymbolic learning are applicable, but in general, the learning for neural programs requires estimating the gradients of black-box components. We present an algorithm for learning neural programs, called ISED, that only relies on input-output samples of black-box components. For evaluation, we introduce new benchmarks that involve calls to modern LLMs such as GPT-4 and also consider benchmarks from the neurosymbolic learning literature. Our evaluation shows that for the latter benchmarks, ISED has comparable performance to state-of-the-art neurosymbolic frameworks. For the former, we use adaptations of prior work on gradient approximations of black-box components as a baseline, and show that ISED achieves comparable accuracy but in a more data- and sample-efficient manner.