Toward Neurosymbolic Program Comprehension

Alejandro Velasco, Aya Garryyeva, David N. Palacio, Antonio Mastropaolo, Denys Poshyvanyk · 2025

Read Toward Neurosymbolic Program Comprehension by Alejandro Velasco, Aya Garryyeva, David N. Palacio, Antonio Mastropaolo, Denys Poshyvanyk (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-01806.txt.

Description

Alejandro Velasco and co-authors argue in 2025 for a hybrid path in AI tools that help programmers understand code. Large code models now assist with writing, testing, and reading software. Yet ever-bigger models cost more to run and stay opaque, which undercuts trust and accountability. The authors question whether scale is always the right road.

Their proposal marries deep learning with classic symbolic methods, long prized for speed and exactness. The paper sketches the core features of such a neurosymbolic framework for program comprehension and shares early results. The target use is spotting faulty parts of a code base, where a wrong guess is costly and a clear reason matters. The team aims to build the first framework of this kind for the task. They frame it as a practical route for organizations that cannot bet on giant models alone.

Catalog record

CREATORAlejandro Velasco, Aya Garryyeva, David N. Palacio, Antonio Mastropaolo, Denys Poshyvanyk
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE21 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-01806

Browse the full library · About the Tower of Babel