Communicating Natural Programs to Humans and Machines

Sam Acquaviva, Yewen Pu, Marta Kryven, Catherine Wong, Theodoros Sechopoulos, Gabrielle Ecanow, Maxwell Nye, Michael Henry Tessler, Joshua B. Tenenbaum

Keywords: program synthesis, communication, cognition

Abstract:

The Abstraction and Reasoning Corpus (ARC) is a set of procedural tasks that tests an agent's ability to flexibly solve novel problems. While most ARC tasks are easy for humans, they are challenging for state-of-the-art AI. What makes building intelligent systems that can generalize to novel situations such as ARC difficult?
We posit that the answer might be found by studying the difference of \emph{language}: While humans readily generate and interpret instructions in a general language, computer systems are shackled to a narrow domain-specific language that they can precisely execute.
We present LARC, the \textit{Language-complete ARC}: a collection of natural language descriptions by a group of human participants who instruct each other on how to solve ARC tasks using language alone, which contains successful instructions for 88\% of the ARC tasks.
We analyze the collected instructions as `natural programs', finding that while they resemble computer programs, they are distinct in two ways: First, they contain a wide range of primitives; Second, they frequently leverage communicative strategies beyond directly executable codes. We demonstrate that these two distinctions prevent current program synthesis techniques from leveraging LARC to its full potential, and give concrete suggestions on how to build the next-generation program synthesizers.

One-sentence Summary: We study natural programs - instructions that can be interpreted by humans to produce a verifiable output, as a pathway to building intelligent systems.

Community Implementations: [ 3 code implementations](https://www.catalyzex.com/paper/arxiv:2106.07824/code)