Neural-Symbolic Learning and Reasoning: A Survey and Interpretation

By Tarek R. Besold, A.S. d'Avila Garcez, Sebastian Bader, H. Bowman, Pedro M. Domingos, Pascal Hitzler, Kai-Uwe Kühnberger, L. C. Lamb, Daniel Lowd, P. Lima and 4 more. 2017

472 citations recorded by the source corpus.

The study and understanding of human behaviour is relevant to computer science, artificial intelligence, neural computation, cognitive science, philosophy, psychology, and several other areas. Presupposing cognition as basis of behaviour, among the most prominent tools in the modelling of behaviour are computational-logic systems, connectionist models of cognition, and models of uncertainty. Recent studies in cognitive science, artificial intelligence, and psychology have produced a number of cognitive models of reasoning, learning, and language that are underpinned by computation. In addition, efforts in computer science research have led to the development of cognitive computational systems integrating machine learning and automated reasoning. Such systems have shown promise in a range of applications, including computational biology, fault diagnosis, training and assessment in simulators, and software verification. This joint survey reviews the personal ideas and views of several researchers on neural-symbolic learning and reasoning. The article is organised in three parts: Firstly, we frame the scope and goals of neural-symbolic computation and have a look at the theoretical foundations. We then proceed to describe the realisations of neural-symbolic computation, systems, and applications. Finally we present the challenges facing the area and avenues for further research.

Subfields: Artificial Intelligence, Computational Theory and Mathematics, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computational Theory and Mathematics

In Natural Language Processing

Citations in the corpus

Learning to Reason Deductively: Math Word Problem Solving as Complex Relation Extraction (2022, background): “MWP solving is typically regarded as one of the system 2 tasks (Kahneman, 2011; Bengio et al., 2021), and our current approach to this problem is related to neural symbolic reasoning (Besold et al., 2017).”
The challenge of crafting intelligible intelligence (2018, methodology): “s are occupied.” While quantified and relational abstractions are very natural for people, progress in statisticalrelationallearninghasbeenslowandtherearemanyopenquestions for neuro-symbolic learning [3]. MappingControlActions:Generating an explanation by mapping an inscrutable model into a simpler,”
Facts as Experts: Adaptable and Interpretable Neural Memory over Symbolic Knowledge (2020, background): “Various neural-symbolic methods have attempted to unify these two extremes (Pinkas, 1991; de Penning et al., 2011; Besold et al., 2017) including many cognitive architectures which used hybrid symbolic and subsymbolic systems (Laird et al., 2017), and more recently, compositional query languages…”
NLProlog: Reasoning with Weak Unification for Question Answering in Natural Language (2019, background): “In Artiﬁcial Intelligence literature, integrating symbolic and sub-symbolic representations is a long-standing problem (Besold et al., 2017).”
Conversational Neuro-Symbolic Commonsense Reasoning (2020, background): “Neuro-symbolic systems are hybrid models that leverage the robustness of connectionist methods and the soundness of symbolic reasoning to effectively integrate learning and reasoning [28,29].”
Interpretable and Trustworthy Deepfake Detection via Dynamic Prototypes (2021, background): “More-over, the difﬁculty in building neuro-symbolic system is the question of how to integrate the output representations of deep networks in a propositional form [7].”
A historical perspective of explainable Artificial Intelligence (2020, background): “As discussed by Besold, Garcez, Bader, et al. (2017), the integration between both levels could, therefore, bridge low-level information processing such as frequently encountered in perception and pattern recognition with reasoning and explanation on a higher, more cognitive level of abstraction.”
Just Add Functions: A Neural-Symbolic Language Model (2019, methodology): “Neural-symbolic integration approaches (Besold et al. 2017) explore general methods of integrating first-order logic with neural networks.”
Explainable AI (XAI): Core Ideas, Techniques, and Solutions (2022, background): “…deep understanding of cognitive science will most certainly beneﬁt the ﬁeld of NS-AI particularly for the integration of neuro-symbolic approaches [6] to develop a mature and robust integrated model and consequently, for the design of explainable systems that are closer to how humans understand,…”
Neural, symbolic and neural-symbolic reasoning on knowledge graphs (2021, background): “…problem of combining symbolism and connectionism has been researched since the 1980s (Gallant, 1988, 1993; Shavlik et al., 1991; Towell and Shavlik, 1994a, 1994b), and has drawn much attention recent years (Besold et al., 2017; de Penning et al., 2011; Garcez et al., 2012, 2019; Lamb et al., 2020).”
From LSAT: The Progress and Challenges of Complex Reasoning (2021, background): “c) Neural-Symbolic Models: To combine the advantages and circumvent the shortcomings of both symbolic and neural methods, neural-symbolic models which integrate symbolic logic and neural representation are widely studied [69, 70, 71, 72, 73].”