cs.AINov 6, 2026

Challenges, Opportunities, and Misconceptions in Neurosymbolic AI

Authors: Giuseppe Marra, Pierre Genevès, Meghyn Bienvenu, Thomas Schiex, Sylvie Thiébaux, Silviu Maniu

Organizations: LPI · CNRS, LIG, TYREX · LaBRI · MIAT INRAE · 

Abstract

We examine opportunities for neurosymbolic AI alongside increasingly powerful neural models, particularly with regards to efficiency, safety and interpretability. We identify two major challenges: ensuring semantic alignment between neural and symbolic components, and managing the computational cost of symbolic inference and learning. We discuss representative methodological approaches to these challenges and address several common misconceptions concerning differentiability, independence assumptions, and scalability. This paper synthesizes perspectives developed through discussions in the Neurosymbolic Programming and Reasoning (NPR) working group at the 2026 CNRS workshop Neurosymbolic Methods and Their Applications in Data Science, rather than providing an exhaustive survey.

Explore similar work

CardsList
  1. The RAIL Principles for Neurosymbolic AI: Reasoning, Assurances, Interfacing and Learning

    Aug 4, 2026Agnese Chiatti, Michael Cochez, Cristina Cornelio +14Symbolic ReasoningNeuro-Symbolic Reasoning

  2. Neuro-Symbolic Artificial Intelligence: A Task-Directed Survey in the Black-Box Models Era

    Mar 3, 2026Giovanni Pio Delvecchio, Lorenzo Molfetta, Gianluca MoroNeuro-Symbolic AINeuro-Symbolic Reasoning

  3. EM-NeSy: Expectation Maximization for Neurosymbolic Learning

    Jun 12, 2026Annegret Seibt, Luc De Raedt, Giuseppe MarraSymbolic ReasoningExpectation-Maximization