Neuro-Symbolic AI in 2024: A Systematic Review

TL;DR

The paper surveys the neuro-symbolic AI landscape from 2020 to 2024 using PRISMA-guided methods, identifying 167 included studies from an initial 1,428 and revealing a strong emphasis on Learning & Inference, Logic & Reasoning, and Knowledge Representation, with notably fewer efforts in Explainability, Trustworthiness, and Meta-Level Cognition. It introduces a five-area Neuro-Symbolic AI Spectrum—Knowledge Representation, Learning and Inference, Explainability and Trustworthiness, Logic and Reasoning—plus a cross-cutting Meta-Level Cognition frontier, and analyzes intersections among these domains. The review highlights AlphaGeometry as a flagship cross-domain project and details cross-area trends, gaps, and opportunities for interdisciplinary work to advance robust, context-aware AI systems. The authors emphasize that addressing gaps in explainability, trustworthiness, and meta-level cognition will be crucial for deploying reliable neuro-symbolic AI in real-world settings and for guiding future research priorities. Overall, the paper provides a taxonomy, a synthesis of recent progress, and a roadmap for integrating symbolic and sub-symbolic approaches to achieve more intelligent and trustworthy AI.

Abstract

Background: The field of Artificial Intelligence has undergone cyclical periods of growth and decline, known as AI summers and winters. Currently, we are in the third AI summer, characterized by significant advancements and commercialization, particularly in the integration of Symbolic AI and Sub-Symbolic AI, leading to the emergence of Neuro-Symbolic AI. Methods: The review followed the PRISMA methodology, utilizing databases such as IEEE Explore, Google Scholar, arXiv, ACM, and SpringerLink. The inclusion criteria targeted peer-reviewed papers published between 2020 and 2024. Papers were screened for relevance to Neuro-Symbolic AI, with further inclusion based on the availability of associated codebases to ensure reproducibility. Results: From an initial pool of 1,428 papers, 167 met the inclusion criteria and were analyzed in detail. The majority of research efforts are concentrated in the areas of learning and inference (63%), logic and reasoning (35%), and knowledge representation (44%). Explainability and trustworthiness are less represented (28%), with Meta-Cognition being the least explored area (5%). The review identifies significant interdisciplinary opportunities, particularly in integrating explainability and trustworthiness with other research areas. Conclusion: Neuro-Symbolic AI research has seen rapid growth since 2020, with concentrated efforts in learning and inference. Significant gaps remain in explainability, trustworthiness, and Meta-Cognition. Addressing these gaps through interdisciplinary research will be crucial for advancing the field towards more intelligent, reliable, and context-aware AI systems.

Figures (3)

• Figure 1: Number of publications per year for neuro-symbolic AI. The data was obtained through Google Scholar scraping, reflecting significant growth from 2020
• Figure 1: The search terms "neurosymbolic" AND each of the terms required for the 5 foundational research areas within neurosymbolic AI were queried through the 5 databases. The number of pieces of literature returned from each query is shown in the table above. Note also that only publications from 2020-2024 were considered
• Figure 2: A literature review of existing of the major components of Symbolic AI was conducted. Note that papers from the Meta-Level Cognition were not required to have an associated public code-base/repository