Neural-Symbolic Computing: An Effective Methodology for Principled Integration of Machine Learning and Reasoning
By A.S. d'Avila Garcez, Marco Gori, L. C. Lamb, Luciano Serafini, Michael Spranger, Son Dinh Tran. 2019
392 citations recorded by the source corpus.
Current advances in Artificial Intelligence and machine learning in general, and deep learning in particular have reached unprecedented impact not only across research communities, but also over popular media channels. However, concerns about interpretability and accountability of AI have been raised by influential thinkers. In spite of the recent impact of AI, several works have identified the need for principled knowledge representation and reasoning mechanisms integrated with deep learning-based systems to provide sound and explainable models for such systems. Neural-symbolic computing aims at integrating, as foreseen by Valiant, two most fundamental cognitive abilities: the ability to learn from the environment, and the ability to reason from what has been learned. Neural-symbolic computing has been an active topic of research for many years, reconciling the advantages of robust learning in neural networks and reasoning and interpretability of symbolic representation. In this paper, we survey recent accomplishments of neural-symbolic computing as a principled methodology for integrated machine learning and reasoning. We illustrate the effectiveness of the approach by outlining the main characteristics of the methodology: principled integration of neural learning with symbolic knowledge representation and reasoning allowing for the construction of explainable AI systems. The insights provided by neural-symbolic computing shed new light on the increasingly prominent need for interpretable and accountable AI systems.
Subfields: Artificial Intelligence, Management Science and Operations Research, Knowledge Graphs & Representation
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
- Francisco Herrera
- Salvador Garcia
- Michael Spranger
- Luciano Serafini
- A.S. d'Avila Garcez
- Alun David Preece
- L. C. Lamb
- Nan Duan
In Management Science and Operations Research
- Elena Hernández-Pereira
- Eduardo Mosqueira-Rey
- Á. Fernández-Leal
- José Bobes-Bascarán
- David Alonso-Ríos
- Svetha Venkatesh
- Tarek R. Besold
- Luciano Serafini
In Knowledge Graphs & Representation
- A.S. d'Avila Garcez
- Luciano Serafini
- Michael Spranger
- L. C. Lamb
- Samy Badreddine
- Adrien Bennetot
- Svetha Venkatesh
- Tarek R. Besold
Citations in the corpus
- Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AI (2019, positive): “Combining connectionist and symbolic paradigms seems a favourable way to address this challenge [169,299,312,352,353] .”
- Rapid Trust Calibration through Interpretable and Uncertainty-Aware AI (2020, background): “92,93 How this can best be achieved is currently a key open problem in AI.”
- 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).”
- A historical perspective of explainable Artificial Intelligence (2020, background): “Last and most importantly in this context, deep learning systems lack transparency while symbolic approaches can be designed in such a way as to follow a humanly comprehensible decision-making process (see, e.g., Garcez et al., 2019; Muggleton, Schmid, Zeller, Tamaddoni-Nezhad, & Besold, 2018).”
- Symbols as a Lingua Franca for Bridging Human-AI Chasm for Explainable and Advisable AI Systems (2021, neutral): “While neuro-symbolic systems have received attention in some quarters (Garcez et al. 2019; De Raedt et al. 2019), the jury is still out on whether or not AI systems need internal symbolic reasoning to reach human-level intelligence.”
- Neurosymbolic AI: the 3rd wave (2020, background): “These include LTN (Serafini and d’Avila Garcez 2016; Stehr et al. 2022) and Tensor Product Representations (Huang et al. 2017), referred to in d’Avila Garcez et al. (2019) as tensorization methods.”
- 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].”
- Right for the Right Concept: Revising Neuro-Symbolic Concepts by Interacting with their Explanations (2020, background): “Neuro-Symbolic architectures [8, 57, 31, 13, 52, 6] make use of data-driven, sub-symbolic representations, and symbol-based reasoning systems.”
- Logic-Driven Context Extension and Data Augmentation for Logical Reasoning of Text (2021, background): “symbolic paradigm (Besold et al., 2017; Garcez et al., 2019) which conducts logic identification and extension in a symbolic manner and utilizes the pre-trained model as the backbone of logic verbalization.”
- Towards Explainable Neural-Symbolic Visual Reasoning (2019, background): “, 2017; Donadello, 2018; d’Avila Garcez et al., 2019]. The use of a Knowledge Base (KB) to learn and reason with symbolic representation has the advantage of promoting the production of explanations while making a prediction [Donadello et al., 2019]. Neural-symbolic computation Manhaeve et al. [2018”
- Neural Reasoning, Fast and Slow, for Video Question Answering (2019, background): “Dual process systems: Reasoning systems that exhibit behaviors consistent with dual process theories are typically neural-symbolic hybrids (e.g., see [8] for an overview).”
- Logic Tensor Networks (2020, background): “Considering also the performance of deep neural networks which have produced state-of-the-art results on image and speech recognition, language modelling, translation and sequence learning, various attempts have been made to combine symbolic ML with neural computation into neural-symbolic systems wi”
- Human-in-the-loop machine learning: a state of the art (2022, background): “Finally, it is worth mentioning that the use of background knowledge in the form of logical statements or constraints in Knowledge Bases (KBs) has shown to not only improve explainability but also performance with respect to purely data-driven approaches (d’Avila Garcez et al. 2019).”