Published Nov 4, 2020

NLP is not NLU and GPT-3 - Walid Saba

Walid Saba challenges the deep learning paradigm, arguing for the existence of innate linguistic structures essential for true language understanding, while critiquing the limitations of AI models in capturing human intelligence and reasoning.

Episode Highlights

Topics covered

Episode Highlights

Related Episodes

#68 DR. WALID SABA 2.0 - Natural Language Understanding [UNPLUGGED]
Answers 383 questions
#88 Dr. WALID SABA - Why machines will never rule the world [UNPLUGGED]
Answers 383 questions
OpenAI GPT-3: Language Models are Few-Shot Learners
Answers 383 questions
#039 - Lena Voita - NLP
Answers 383 questions
#031 WE GOT ACCESS TO GPT-3! (With Gary Marcus, Walid Saba and Connor Leahy)
Answers 383 questions
Explainability, Reasoning, Priors and GPT-3
Answers 383 questions
#53 Quantum Natural Language Processing - Prof. Bob Coecke (Oxford)
Answers 383 questions
#029 GPT-3, Prompt Engineering, Trading, AI Alignment, Intelligence
Answers 383 questions
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
Answers 383 questions
#56 - Dr. Walid Saba, Gadi Singer, Prof. J. Mark Bishop (Panel discussion)
Answers 383 questions
Jürgen Schmidhuber - Neural and Non-Neural AI, Reasoning, Transformers, and LSTMs
Answers 383 questions
#83 Dr. ANDREW LAMPINEN (Deepmind) - Natural Language, Symbols and Grounding [NEURIPS2022 UNPLUGGED]
Answers 383 questions
UK Algoshambles, Neuralink, GPT-3 and Intelligence
Answers 383 questions
#84 LAURA RUIS - Large language models are not zero-shot communicators [NEURIPS UNPLUGGED]
Answers 383 questions
MLST #78 - Prof. NOAM CHOMSKY (Special Edition)
Answers 383 questions

Dexa/Machine Learning Street Talk (MLST)

NLP is not NLU and GPT-3 - Walid Saba

Topics covered

Popular Clips

Intelligence, Computation

Natural Language Understanding

Evolution of NLP

Ontological Concepts in Neural Networks

Knowledge Graphs and Data

Flexibility in Type Systems

Language Structure Insights

Word Embeddings Insight

AI Evolution Insights

Paradigm Shift

Natural Language Reasoning

Ambiguity in Language

Evolution of Deep Learning

Unraveling Language Compiler

GPT-3 Critique

Episode Highlights

Concepts of Intelligence

Natural Language Understanding

Deep Learning Limitations

Related Episodes

#68 DR. WALID SABA 2.0 - Natural Language Understanding [UNPLUGGED]

#88 Dr. WALID SABA - Why machines will never rule the world [UNPLUGGED]

OpenAI GPT-3: Language Models are Few-Shot Learners

#039 - Lena Voita - NLP

#031 WE GOT ACCESS TO GPT-3! (With Gary Marcus, Walid Saba and Connor Leahy)

Explainability, Reasoning, Priors and GPT-3

#53 Quantum Natural Language Processing - Prof. Bob Coecke (Oxford)

#029 GPT-3, Prompt Engineering, Trading, AI Alignment, Intelligence

Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer

#56 - Dr. Walid Saba, Gadi Singer, Prof. J. Mark Bishop (Panel discussion)

Jürgen Schmidhuber - Neural and Non-Neural AI, Reasoning, Transformers, and LSTMs

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

UK Algoshambles, Neuralink, GPT-3 and Intelligence

#84 LAURA RUIS - Large language models are not zero-shot communicators [NEURIPS UNPLUGGED]

MLST #78 - Prof. NOAM CHOMSKY (Special Edition)

NLP is not NLU and GPT-3 - Walid Saba

Topics covered

Popular Clips

Episode Highlights

Concepts of IntelligenceThe conversation shifts to cognitive plausibility and the nature of intelligence, questioning the reliance on computational power in AI models. Walid Saba and Tim Scarfe explore the need for innate knowledge structures to truly understand language.

Concepts of Intelligence

Natural Language UnderstandingWalid Saba presents a compelling argument for the existence of innate linguistic structures that aid in language comprehension. He challenges the deep learning paradigm, advocating for a more nuanced understanding of language as a formal system with inherent rules.

Natural Language Understanding

Deep Learning LimitationsWalid Saba critiques deep learning's limitations in capturing human reasoning and intelligence. He argues for a reevaluation of empirical methods and highlights the structural shortcomings of neural networks.

Deep Learning Limitations

Related Episodes