Machine Learning Street Talk (MLST) artwork

#83 Dr. ANDREW LAMPINEN (Deepmind) - Natural Language, Symbols and Grounding [NEURIPS2022 UNPLUGGED]

Machine Learning Street Talk (MLST)

English - December 04, 2022 07:51 - 20 minutes - 47.2 MB
Technology Homepage Download Google Podcasts Overcast Castro Pocket Casts RSS feed
Previous Episode: #82 - Dr. JOSCHA BACH - Digital Physics, DL and Consciousness [UNPLUGGED]
Next Episode: #84 LAURA RUIS - Large language models are not zero-shot communicators [NEURIPS UNPLUGGED]

First in our unplugged series live from #NeurIPS2022


We discuss natural language understanding, symbol meaning and grounding and Chomsky with Dr. Andrew Lampinen from DeepMind. 


We recorded a LOT of material from NeurIPS, keep an eye out for the uploads. 




YT version: https://youtu.be/46A-BcBbMnA




References


[Paul Cisek] Beyond the computer metaphor: Behaviour as interaction


https://philpapers.org/rec/CISBTC




Linguistic Competence (Chomsky reference)


https://en.wikipedia.org/wiki/Linguistic_competence




[Andrew Lampinen] Can language models handle recursively nested grammatical structures? A case study on comparing models and humans


https://arxiv.org/abs/2210.15303




[Fodor et al] Connectionism and Cognitive Architecture: A Critical Analysis


https://ruccs.rutgers.edu/images/personal-zenon-pylyshyn/proseminars/Proseminar13/ConnectionistArchitecture.pdf




[Melanie Mitchell et al] The Debate Over Understanding in AI's Large Language Models


https://arxiv.org/abs/2210.13966




[Gary Marcus] GPT-3, Bloviator: OpenAI’s language generator has no idea what it’s talking about


https://www.technologyreview.com/2020/08/22/1007539/gpt3-openai-language-generator-artificial-intelligence-ai-opinion/




[Bender et al] On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?


https://dl.acm.org/doi/10.1145/3442188.3445922




[Adam Santoro, Andrew Lampinen et al] Symbolic Behaviour in Artificial Intelligence


https://arxiv.org/abs/2102.03406




[Ishita Dasgupta, Lampinen et al] Language models show human-like content effects on reasoning


https://arxiv.org/abs/2207.07051




REACT - Synergizing Reasoning and Acting in Language Models


https://arxiv.org/pdf/2210.03629.pdf


https://ai.googleblog.com/2022/11/react-synergizing-reasoning-and-acting.html




[Fabian Paischer] HELM - History Compression via Language Models in Reinforcement Learning


https://ml-jku.github.io/blog/2022/helm/


https://arxiv.org/abs/2205.12258




[Laura Ruis] Large language models are not zero-shot communicators


https://arxiv.org/pdf/2210.14986.pdf




[Kumar] Using natural language and program abstractions to instill human inductive biases in machines


https://arxiv.org/pdf/2205.11558.pdf




Juho Kim


https://juhokim.com/