Knowledge Tree
Gary Marcus
Facts (11)
Sources
Neuro-symbolic AI - Wikipedia en.wikipedia.org 5 facts
claimGary Marcus identifies four cognitive prerequisites for building robust artificial intelligence: (1) hybrid architectures that combine large-scale learning with the representational and computational powers of symbol manipulation, (2) large-scale knowledge bases—likely leveraging innate frameworks—that incorporate symbolic knowledge along with other forms of knowledge, (3) reasoning mechanisms capable of leveraging those knowledge bases in tractable ways, and (4) rich cognitive models that work together with those mechanisms and knowledge bases.
quoteGary Marcus stated: "To build a robust, knowledge-driven approach to AI we must have the machinery of symbol manipulation in our toolkit. Too much of useful knowledge is abstract to make do without tools that represent and manipulate abstraction, and to date, the only known machinery that can manipulate such abstract knowledge reliably is the apparatus of symbol manipulation."
referenceGary Marcus and Ernest Davis authored the book 'Rebooting AI: Building Artificial Intelligence We Can Trust', which discusses the development of trustworthy artificial intelligence.
quoteGary Marcus stated: "We cannot construct rich cognitive models in an adequate, automated way without the triumvirate of hybrid architecture, rich prior knowledge, and sophisticated techniques for reasoning."
perspectiveGary Marcus argues that hybrid architectures combining learning and symbol manipulation are necessary but not sufficient for robust artificial intelligence.
Unlocking the Potential of Generative AI through Neuro-Symbolic ... arxiv.org Feb 16, 2025 3 facts
referenceGary Marcus published the preprint 'Deep learning: A critical appraisal' on arXiv in 2018.
claimThe 2019 Montreal AI Debate between Gary Marcus and Yoshua Bengio catalyzed a surge of interest in hybrid neuro-symbolic artificial intelligence solutions by highlighting the contrasting perspectives on the future of AI.
quoteGary Marcus argued during the 2019 Montreal AI Debate that 'expecting a monolithic architecture to handle abstraction and reasoning is unrealistic,' emphasizing the limitations of current AI systems.
Neurosymbolic AI: The Future of AI After LLMs - LinkedIn linkedin.com Nov 11, 2025 1 fact
claimGary Marcus has been advocating for a pivot toward neurosymbolic AI.
The Synergy of Symbolic and Connectionist AI in LLM-Empowered ... arxiv.org Jul 11, 2024 1 fact
claimYann LeCun, Yoshua Bengio, and Gary Marcus have engaged in historical debates that underscore the limitations of both connectionist and symbolic AI approaches.
Consciousness in Artificial Intelligence? A Framework for Classifying ... arxiv.org Nov 20, 2025 1 fact
claimJosh Tenenbaum, Gary Marcus, Emily Bender, Alexander Koller, and Yann LeCun have developed arguments suggesting that specific current algorithms are missing crucial elements required to achieve general intelligence or understanding.