Machine Learning Street Talk (MLST)

chevron_right

#74 Dr. ANDREW LAMPINEN - Symbolic behaviour in AI [UNPLUGGED]

Apr 14, 2022

Guest

Andrew Lampinen

Dr. Andrew Lampinen, a Senior Research Scientist at DeepMind with a PhD from Stanford, dives into the nuanced world of symbolic behavior in AI. He discusses how machines struggle to replicate human symbol use and emphasizes that meanings are shaped by user agreements rather than symbol content. Lampinen critiques traditional intelligence notions, advocating for a meaning-first approach in AI. The conversation also touches on the complexity of subjectivity, the limits of formal logic, and the ethical challenges in aligning AI with human values.

01:05:38

forum

Ask episode

web_stories

AI Snips

view_agenda

Chapters

menu_book

Books

auto_awesome

Transcript

info_circle

Episode notes

insights

INSIGHT

Semantic Intuition

Humans use semantic intuition as a heuristic.
This focus helps navigate the vast space of symbolic manipulation, making it more efficient.

insights

INSIGHT

Subjectivity in Interpretation

Subjectivity is a complex process entangled with the interpreter's state.
Even seemingly objective frameworks contain subjectivity in how systems are set up.

insights

INSIGHT

Syllogisms as Abstractions

Syllogisms and logic rules are useful abstractions describing observed patterns, like Newtonian mechanics.
Whether they are fundamental or just useful is unclear, especially without an objective definition of truth.

Get the Snipd Podcast app to discover more snips from this episode

Symbolic Behavior and AI: A Meaning-First Approach

02:12 • 11min

chevron_right

The Complexity of Subjectivity

12:56 • 8min

chevron_right

Exploring Universal Cognitive Templates and Language in Intelligence

21:04 • 2min

chevron_right

Logic and Reasoning in AI

23:14 • 8min

chevron_right

Language, Intelligence, and AI Ethics

30:52 • 33min

chevron_right

Exploring Intelligence and Consciousness in AI

01:03:54 • 2min

chevron_right

#2187

• Mentioned in 20 episodes

Philosophical Investigations

Ludwig Wittgenstein

Published posthumously in 1953, 'Philosophical Investigations' is a seminal work by Ludwig Wittgenstein that challenges many of the ideas presented in his earlier work, 'Tractatus Logico-Philosophicus'. The book is divided into two parts and consists of short observations or 'remarks' that delve into topics such as the theory of language, language games, meaning, symbols, concepts, and categories. Wittgenstein argues that the meaning of a word is derived from its use within the context of a language-game, rejecting the idea that words gain meaning by referencing objects or mental representations. He emphasizes the importance of understanding language as a tool for communication and social interaction, rather than as a system for representing objective reality.

#1472

• Mentioned in 28 episodes

So You've Been Publicly Shamed

Jon Ronson

In 'So You've Been Publicly Shamed', Jon Ronson delves into the phenomenon of online shaming and its historical antecedents. The book features interviews with individuals who have been subjected to intense public shaming, including Jonah Lehrer, a journalist who plagiarized quotes, and a woman who was shamed for a tweet at an airport. Ronson also speaks with practitioners of modern public humiliation and examines the immediate and long-term consequences of such shaming. The book raises questions about righteousness, reputation, and conformity in the context of social media and its impact on individuals and society.

#15073

• Mentioned in 3 episodes

The Philosophy of Physical Science

Sir Arthur Eddington

Sir Arthur Eddington's "The Philosophy of Physical Science" is a classic work exploring the relationship between physics and philosophy. Eddington delves into the fundamental concepts of physics, examining the nature of space, time, and causality. He challenges traditional interpretations of scientific knowledge, emphasizing the role of observation and the limitations of human perception. Eddington's work is known for its insightful discussions on the nature of scientific laws and their relationship to reality. He explores the philosophical implications of relativity and quantum mechanics, offering a unique perspective on the nature of the universe. The book remains a significant contribution to the philosophy of science, inspiring further discussions on the limits of scientific knowledge and the role of interpretation in scientific understanding.

Please note that in this interview Dr. Lampinen was expressing his personal opinions and they do not necessarily represent those of DeepMind.

Patreon: https://www.patreon.com/mlst

Discord: https://discord.gg/ESrGqhf5CB

YT version: https://youtu.be/yPMtSXXn4OY

Dr. Andrew Lampinen is a Senior Research Scientist at DeepMind, and he thinks that symbols are subjective in the relativistic sense. Dr. Lampinen completed his PhD in Cognitive Psychology at Stanford University. His background is in mathematics, physics, and machine learning. Andrew has said that his research interests are in cognitive flexibililty and generalization, and how these abilities are enabled by factors like language, memory, and embodiment. Andrew with his coauthors has just released a paper called symbolic behaviour in artificial intelligence. Andrew lead in the paper by saying the human ability to use symbols has yet to be replicated in machines. He thinks that one of the key areas to bridge the gap here is considering how symbol meaning is established, and he strongly believes it is the symbol users themselves who agree upon the symbol meaning, And that the use of symbols entails behaviours which coalesce agreements about their meaning. Which in plain English means that symbols are defined by behaviours rather than their content.

[00:00:00] Intro to Andrew and Symbolic Behaviour paper

[00:07:01] Semantics underpins the unreasonable effectiveness of symbols

[00:12:56] The Depth of Subjectivity

[00:21:03] Walid Saba - universal cognitive templates

[00:27:47] Insufficiently Darwinian

[00:30:52] Discovered vs invented

[00:34:19] Does language have primacy

[00:35:59] Research directions

[00:39:43] Comparison to BenG OpenCog and human compatible AI

[00:42:53] Aligning AI with our culture

[00:47:55] Do we need to model the worst aspects of human behaviour?

[00:50:57] Fairness

[00:54:24] Memorisatation on LLMs

[01:00:38] Wason selection task

[01:03:45] Would an Andrew hashtable robot be intelligent?

Dr. Andrew Lampinen

https://lampinen.github.io/

https://twitter.com/AndrewLampinen

Symbolic Behaviour in Artificial Intelligence

https://arxiv.org/abs/2102.03406

Imitating Interactive Intelligence

https://arxiv.org/abs/2012.05672

https://www.deepmind.com/publications/imitating-interactive-intelligence

Impact of Pretraining Term Frequencies on Few-Shot Reasoning [Yasaman Razeghi]

https://arxiv.org/abs/2202.07206

Big bench dataset

https://github.com/google/BIG-bench

Teaching Autoregressive Language Models Complex Tasks By Demonstration [Recchia]

https://arxiv.org/pdf/2109.02102.pdf

Wason selection task

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

Gary Lupyan

https://psych.wisc.edu/staff/lupyan-gary/

Home Top podcasts Popular guests Top books