Knowledge Tree
logical reasoning
Also known as: logical thinking
Facts (27)
Sources
Neuro-Symbolic AI: Explainability, Challenges, and Future Trends arxiv.org Nov 7, 2024 7 facts
referenceSATNet, developed by Po-Wei Wang, Priya Donti, Bryan Wilder, and Zico Kolter in 2019, is a differentiable satisfiability solver designed to bridge deep learning and logical reasoning.
referenceLiangming Pan, Alon Albalak, Xinyi Wang, and William Yang Wang introduced Logic-LM, a framework for empowering large language models with symbolic solvers for faithful logical reasoning, in 2023.
referenceNeuro-symbolic AI research is categorized into several domains: mathematics and symbolic regression (e.g., Majumdar et al., 2023; Petersen et al., 2019), logic and knowledge processing including concept/rule learning (e.g., Aspis et al., 2022) and logical reasoning (e.g., Cunnington et al., 2023), and applications such as visual question answering (e.g., Mao et al., 2019), medical (e.g., Jain et al., 2023), communication (e.g., Thomas and Saad, 2023), programming (e.g., Hu et al., 2022a), recommendation systems (e.g., Carraro, 2023), and security (e.g., Wang et al., 2018).
claimLogical Neural Networks (LNN) can simultaneously perform multiple logical reasoning tasks, such as theorem proving and fact derivation, unlike traditional single-task neural networks.
perspectiveResearch into brain function, such as the separation of logical reasoning from language processing, may inspire the development of more flexible reasoning paths or dynamically configurable task reasoning methods in Neuro-Symbolic AI.
referenceCoetzee et al. (2022) found that logical reasoning in the adult brain may be separated from language processing.
claimProcess transparency in Neuro-Symbolic AI requires that the generation of symbols for logical reasoning by neural networks be transparent and interpretable enough to verify correctness, potentially through rigorous logic or formulaic arguments.
Unlocking the Potential of Generative AI through Neuro-Symbolic ... arxiv.org Feb 16, 2025 5 facts
claimSymbolic[Neuro], Neuro Symbolic Neuro, and NeuroSymbolicNeuro architectures show strong capabilities in logical reasoning and relational understanding.
claimReasoning evaluation in Neuro-Symbolic AI (NSAI) architectures includes logical reasoning (systematic application of explicit rules for inferences), relational understanding (comprehension of complex relationships between entities), and cognitive versatility (integration of deductive, inductive, and abductive reasoning paradigms).
referenceThe paper 'RAG-logic: Enhance neuro-symbolic approaches for logical reasoning with retrieval-augmented generation' by Anonymous was submitted to the ACL Rolling Review in June 2024.
referenceZeming Chen, Qiyue Gao, and Lawrence S. Moss developed NeuralLog, a system for natural language inference using joint neural and logical reasoning, published as an arXiv preprint in 2021.
procedureRAG-Logic is a dynamic example-based framework that enhances logical reasoning by integrating contextually appropriate examples using a four-step process: (1) encoding symbolic input into neural representations via a RAG knowledge base search module, (2) transforming input into formal logical formulas via a translation module, (3) ensuring syntactic correctness via a fix module, and (4) evaluating logical consistency via a solver module before decoding results back into symbolic output.
Neuro-symbolic AI - Wikipedia en.wikipedia.org 3 facts
claimAbductive Learning integrates machine learning and logical reasoning in a balanced-loop via abductive reasoning, enabling the two approaches to work together in a mutually beneficial way.
referenceLuciano Serafini and Artur d'Avila Garcez authored the 2016 paper 'Logic Tensor Networks: Deep Learning and Logical Reasoning from Data and Knowledge', published on arXiv.
referenceLuciano Serafini and Artur d'Avila Garcez authored 'Logic Tensor Networks: Deep Learning and Logical Reasoning from Data and Knowledge', which discusses integrating deep learning with logical reasoning.
The Synergy of Symbolic and Connectionist AI in LLM ... arxiv.org 2 facts
claimThe integration of connectionist and symbolic paradigms has led to hybrid models that combine the pattern recognition capabilities of neural networks with the interpretability and logical reasoning of symbolic systems.
imageFigure 4 in the paper illustrates the P2oT framework verifying a mathematical proof that the sum of the first n positive integers is an even number for any even integer n, using the Dafny tool for formal verification to demonstrate the framework's structured approach to logical reasoning.
A survey on augmenting knowledge graphs (KGs) with large ... link.springer.com Nov 4, 2024 2 facts
claimReClor is a benchmark for evaluating logical reasoning in reading comprehension by testing a model's ability to understand and reason through logical relationships in text.
referenceYu W, Jiang Z, Dong Y, and Feng J authored 'Reclor: A reading comprehension dataset requiring logical reasoning', published as an arXiv preprint in 2020 (arXiv:2002.04326).
The Synergy of Symbolic and Connectionist AI in LLM-Empowered ... arxiv.org Jul 11, 2024 1 fact
claimHybrid AI models integrate connectionist AI's pattern recognition with symbolic AI's interpretability and logical reasoning to create more robust systems.
Construction of Knowledge Graphs: State and Challenges - arXiv arxiv.org 1 fact
claimThe traditional method for determining missing type information in RDF datasets involves logical reasoning, but this approach is limited because it relies on already consistent facts and existing rdf:type information within the knowledge base.
A Survey of Incorporating Psychological Theories in LLMs - arXiv arxiv.org 1 fact
referenceNan Xu, Fei Wang, Ben Zhou, Bangzheng Li, Chaowei Xiao, and Muhao Chen authored 'Cognitive overload: Jailbreaking large language models with overloaded logical thinking', published in the Findings of the Association for Computational Linguistics: NAACL 2024.
KG-IRAG: A Knowledge Graph-Based Iterative Retrieval-Augmented ... arxiv.org Mar 18, 2025 1 fact
referenceThe KG-IRAG evaluation uses three question types: Q1, which is a fundamental entity recognition and retrieval task; and Q2 and Q3, which introduce logical reasoning by incorporating time-dependent queries to test iterative reasoning over time.
Evolutionary psychology - Wikipedia en.wikipedia.org 1 fact
referencePaul Sheldon Davies, James H. Fetzer, and Thomas R. Foster published 'Logical reasoning and domain specificity: A critique of the social exchange theory of reasoning' in Biology & Philosophy in 1995.
Neural-Symbolic AI: The Next Breakthrough in Reliable and ... hu.ac.ae Dec 29, 2025 1 fact
claimNeuro-symbolic AI in self-driving cars merges perception with logical reasoning, allowing vehicles to enhance image recognition and apply rule-based reasoning to minimize errors in unpredictable situations.
Cybersecurity Trends and Predictions 2025 From Industry Insiders itprotoday.com 1 fact
claimNeuro-Symbolic AI (NSAI) will combine pattern recognition, logical reasoning, and language understanding to identify suspicious transactions across decentralized platforms, helping regulators and industry players maintain transparency and compliance.
Construction of intelligent decision support systems through ... - Nature nature.com Oct 10, 2025 1 fact
claimStructured reasoning methods, such as SMT solvers, perform well on formal logical reasoning but struggle with natural language interaction and ambiguous query situations.