Defining neurosymbolic AI

Neurosymbolic AI focuses on integrating learning and reasoning, in particular, on unifying logical and neural representations. Despite the existence of an alphabet soup of neurosymbolic AI systems, the field is lacking a generally accepted formal definition of what neurosymbolic models and inference really are. We introduce a formal definition for neurosymbolic AI that makes abstraction of its key ingredients. More specifically, we define neurosymbolic inference as the computation of an integral over a product of a logical and a belief function. We show that our neurosymbolic AI definition makes abstraction of key representative neurosymbolic AI systems.

• Publication year

  2025

• Venue

  arXiv.org

• Publication date

  2025-07-15

• Fields of study

  Philosophy, Computer Science

• Identifiers
  DOI 10.48550/arXiv.2507.11127arXiv 2507.11127
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Beyond the convexity assumption: Realistic tabular data generation under quantifier-free real linear constraints

  2025cited by this paper
• Π-NeSy: A Possibilistic Neuro-Symbolic Approach

  2025cited by this paper
• Valid Text-to-SQL Generation with Unification-Based DeepStochLog

  2025influential reference
• A semantic framework for neurosymbolic computation

  2024cited by this paper
• NeSyA: Neurosymbolic Automata

  2024cited by this paper
• ULLER: A Unified Language for Learning and Reasoning

  2024cited by this paper
• Logic of Differentiable Logics: Towards a Uniform Semantics of DL

  2023cited by this paper
• Neural Probabilistic Logic Programming in Discrete-Continuous Domains

  2023cited by this paper
• $$\alpha$$ α ILP: thinking visual scenes as differentiable logic programs

  2023cited by this paper
• Safe Reinforcement Learning via Probabilistic Logic Shields

  2023cited by this paper
• Declarative Probabilistic Logic Programming in Discrete-Continuous Domains

  2023cited by this paper
• Soft-Unification in Deep Probabilistic Logic

  2023cited by this paper
• Neurosymbolic AI for Reasoning on Biomedical Knowledge Graphs

  2023cited by this paper
• Scallop: A Language for Neurosymbolic Programming

  2023cited by this paper
• Semantic Probabilistic Layers for Neuro-Symbolic Learning

  2022cited by this paper
• The Third AI Summer: AAAI Robert S. Engelmore Memorial Lecture

  2022cited by this paper
• Toward a broad AI

  2022cited by this paper
• Neurosymbolic Motion and Task Planning for Linear Temporal Logic Tasks

  2022cited by this paper
• Visual Sudoku Puzzle Classification: A Suite of Collective Neuro-Symbolic Tasks

  2022influential reference
• Neural-Probabilistic Answer Set Programming

  2022cited by this paper
• NeuPSL: Neural Probabilistic Soft Logic

  2022cited by this paper
• Compositional Approach to Translate LTLf/LDLf into Deterministic Finite Automata

  2021cited by this paper
• Neuro-Symbolic Artificial Intelligence: The State of the Art

  2021cited by this paper
• From statistical relational to neurosymbolic artificial intelligence: A survey

  2021cited by this paper
• How to Tell Deep Neural Networks What We Know

  2021cited by this paper
• DeepStochLog: Neural Stochastic Logic Programming

  2021cited by this paper
• Modular design patterns for hybrid learning and reasoning systems

  2021cited by this paper
• From Statistical Relational to Neuro-Symbolic Artificial Intelligence

  2020cited by this paper
• TensorLog: A Probabilistic Database Implemented Using Deep-Learning Infrastructure

  2020influential reference
• Probabilistic Circuits: A Unifying Framework for Tractable Probabilistic Models∗

  2020cited by this paper
• Neural-Symbolic Integration: A Compositional Perspective

  2020cited by this paper
• Neurosymbolic AI: the 3rd wave

  2020cited by this paper
• Logic Tensor Networks

  2020cited by this paper
• NeurASP: Embracing Neural Networks into Answer Set Programming

  2020cited by this paper
• Neural Markov Logic Networks

  2019cited by this paper
• A Boxology of Design Patterns for Hybrid Learning and Reasoning Systems

  2019cited by this paper
• Neural Logic Machines

  2019cited by this paper
• NLProlog: Reasoning with Weak Unification for Question Answering in Natural Language

  2019cited by this paper
• Integrating Learning and Reasoning with Deep Logic Models

  2019cited by this paper
• Conditional Sum-Product Networks: Imposing Structure on Deep Probabilistic Architectures

  2019cited by this paper
• Neural-Symbolic Computing: An Effective Methodology for Principled Integration of Machine Learning and Reasoning

  2019cited by this paper
• Neural probabilistic logic programming in DeepProbLog

  2019cited by this paper
• DL2: Training and Querying Neural Networks with Logic

  2019cited by this paper
• DeepProbLog: Neural Probabilistic Logic Programming

  2018cited by this paper
• Lifted Relational Neural Networks: Efficient Learning of Latent Relational Structures

  2018cited by this paper
• Inductive Learning of Answer Set Programs from Noisy Examples

  2018cited by this paper
• A Semantic Loss Function for Deep Learning with Symbolic Knowledge

  2017influential reference
• Differentiable Learning of Logical Rules for Knowledge Base Reasoning

  2017cited by this paper
• Neural-Symbolic Learning and Reasoning: A Survey and Interpretation

  2017cited by this paper
• Semantic-based regularization for learning and inference

  2017cited by this paper
• End-to-end Differentiable Proving

  2017cited by this paper
• Tensor Networks : Deep Learning and Logical Reasoning from Data and Knowledge ⋆

  2016cited by this paper
• Statistical Relational Artificial Intelligence: Logic, Probability, and Computation

  2016cited by this paper
• Probabilistic Inference in Hybrid Domains by Weighted Model Integration

  2015cited by this paper
• Linear Temporal Logic and Linear Dynamic Logic on Finite Traces

  2013cited by this paper
• Algebraic model counting

  2012cited by this paper
• On probabilistic inference by weighted model counting

  2008cited by this paper
• Detecting change in vague interpretations of landscapes

  2006cited by this paper
• Dimensions of Neural-symbolic Integration - A Structured Survey

  2005cited by this paper
• Fuzzy sets as a basis for a theory of possibility

  1999cited by this paper
• An Automata-Theoretic Approach to Linear Temporal Logic

  1996cited by this paper
• Theory of T -norms and fuzzy inference methods

  1991cited by this paper
• On the semantics of fuzzy logic

  1991cited by this paper
• A Brief Overview of Possibilistic Logic

  1991cited by this paper
• The Stable Model Semantics for Logic Programming

  1988cited by this paper
• Fuzzy Measure and Fuzzy Integral

  1972cited by this paper
• Theory of capacities

  1954cited by this paper

• The DeepLog Neurosymbolic Machine

  2025influential citation
• CLAUSE: Agentic Neuro-Symbolic Knowledge Graph Reasoning via Dynamic Learnable Context Engineering

  2025cites this paper
• Symbolic Neural Generation with Applications to Lead Discovery in Drug Design

  2025cites this paper
• Investigating Neurosymbolic AI for Intent-based Service Management

  2025cites this paper