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/