Gary Marcus
In plain wordsAn American cognitive scientist and leading critic of deep learning, who argues that today's AI cannot reason reliably and needs to be combined with symbolic methods.
American cognitive scientist and a leading critic of deep learningdeep learning, n.: Machine learning done with neural networks built from many stacked layers: raw data such as pixels, sound or text goes in at one end, and each layer turns what it receives into something more abstract. as a route to general intelligence; he is an emeritus professor at New York University, where he taught psychology and neural science. Trained under Steven Pinker, he studied how children learn grammatical rules, and in 'The Algebraic Mind' (2001) argued that neural networks without explicit symbols cannot reliably generalize abstract rules beyond their training data. He applied the argument to modern systems in essays such as 'Deep Learning: A Critical Appraisal' (2018) and 'Deep Learning Is Hitting a Wall' (2022) and in 'Rebooting AI' (2019, with Ernest Davis), pointing to failures of reasoning, hallucinationhallucination, n.: When an AI states something false or made up — a fake fact, quotation or source — fluently and confidently, as if it were true. and out-of-distributionout-of-distribution, adj.: Describes inputs that differ from anything an AI system saw in training — like a model trained on daytime photos facing a night scene — where its answers become less reliable. generalization, and advocates hybrid neuro-symbolic AIneuro-symbolic AI, n.: AI that combines neural networks, which learn patterns from data, with symbolic methods, which follow explicit rules and logic, aiming for systems that both learn flexibly and reason reliably.. He founded the startup Geometric Intelligence, acquired by Uber in 2016, and co-founded Robust.AI with Rodney BrooksRodney Brooks, prop. n.: An Australian-American roboticist who argued that intelligence grows out of acting in the world rather than abstract reasoning, and who co-founded the company behind the Roomba vacuum.. In May 2023 he testified to a US Senate subcommittee alongside Sam AltmanSam Altman, prop. n.: An American entrepreneur who co-founded OpenAI and runs it; the company's 2022 release of ChatGPT turned him into one of the public faces of the AI era., urging AI regulation, a case he developed in 'Taming Silicon Valley' (2024).
Related
- neuro-symbolic AIneuro-symbolic AI, n.: AI that combines neural networks, which learn patterns from data, with symbolic methods, which follow explicit rules and logic, aiming for systems that both learn flexibly and reason reliably.
- deep learningdeep learning, n.: Machine learning done with neural networks built from many stacked layers: raw data such as pixels, sound or text goes in at one end, and each layer turns what it receives into something more abstract.
- AI hypeAI hype, n.: Exaggerated claims about what AI can do or will soon do, made to attract money, attention or customers, which can mislead the public and policymakers.
- hallucinationhallucination, n.: When an AI states something false or made up — a fake fact, quotation or source — fluently and confidently, as if it were true.
- out-of-distributionout-of-distribution, adj.: Describes inputs that differ from anything an AI system saw in training — like a model trained on daytime photos facing a night scene — where its answers become less reliable.
- symbolic AIsymbolic AI, n.: An approach to AI that writes knowledge down as human-readable symbols and rules, such as 'every bird has wings', and reasons by manipulating them step by step instead of learning patterns from data.
Sources
- Marcus, G. F. (2001). The Algebraic Mind: Integrating Connectionism and Cognitive Science. MIT Press.
- Marcus, G. (2018). Deep Learning: A Critical Appraisal. arXiv preprint.
Mentioned in
neuro-symbolic AIneuro-symbolic AI, n.: AI that combines neural networks, which learn patterns from data, with symbolic methods, which follow explicit rules and logic, aiming for systems that both learn flexibly and reason reliably., scaling wallscaling wall, n.: The disputed idea that making AI models bigger is running out of steam, so progress will soon stall; 'hitting a wall' became the skeptics' slogan and 'there is no wall' the reply., Douglas LenatDouglas Lenat, prop. n.: An American computer scientist who spent four decades building Cyc, a giant hand-written database of everyday common sense meant to let computers reason like people.