Learning, Fast and Slow: 10 Years Plan - Memory Soup, Hier. Planning, Emotions, Knowledge Sharing

Subscribers:
5,350
Published on ● Video Link: https://www.youtube.com/watch?v=edYOGUZkFpY



Duration: 2:31:49
603 views
35


How can agents learn and adapt fast in changing environments?

Neural network weights alone are too slow, we need memory.

How can memory be encoded and retrieved, how can goals be decided and abstracted? What role will emotions play? How does memory soup, where all levels of abstraction go into one common soup, fit in?

This is a follow-up to my ICDL 2023 paper "Learning, Fast and Slow", which was awarded Best Paper Finalist.

There are more questions than answers, but solving them will ultimately lead us closer to the end goal of more adaptive systems, which could have numerous use-cases we have yet to even think of yet.

~~~~~~~~~~~~~~
Slides: https://github.com/tanchongmin/TensorFlow-Implementations/blob/main/Paper_Reviews/Learning%2C%20Fast%20and%20Slow%20(10%20years%20plan).pdf

Base Learning, Fast and Slow Model (15 min recap): https://www.youtube.com/watch?v=DSVFA7nmwHQ
Code: https://github.com/tanchongmin/learning-fast-and-slow
Paper: https://arxiv.org/abs/2301.13758

~~~~~~~~~~~~~~

0:00 Introduction
1:10 Base Model Recap
42:24 10 Year Plan Overview
50:43 Goal-Directed Decision Transformer
52:10 Hierarchical Prediction (Yann LeCun)
54:48 Hierarchical Prediction and Multiple Abstraction Spaces (Fast and Slow)
1:15:42 Curriculum Learning to add to Self-Learning
1:18:33 How to Define Goals
1:25:40 Building Abstracted Goals by Proximity
1:34:20 Different Embedding Spaces in Same Memory Soup
1:37:13 Memory Soup
1:41:56 How to Learn Different Abstraction Spaces by Self-Supervised Learning
1:52:10 Emotion-Guided Memory Retrieval
2:02:02 Knowledge Sharing Between Agents
2:14:50 Multimodal Memory Search
2:21:53 Can everything be modelled as a graph traversal problem?
2:28:13 Fast and Slow 10 Year Plan Overall Recap and Conclusion

~~~~~~~~~~~~~~

AI and ML enthusiast. Likes to think about the essences behind breakthroughs of AI and explain it in a simple and relatable way. Also, I am an avid game creator.

Discord: https://discord.gg/bzp87AHJy5
LinkedIn: https://www.linkedin.com/in/chong-min-tan-94652288/
Online AI blog: https://delvingintotech.wordpress.com/
Twitter: https://twitter.com/johntanchongmin
Try out my games here: https://simmer.io/@chongmin




Other Videos By John Tan Chong Min


2024-02-27SymbolicAI / ExtensityAI Paper Overview (Part 2) - Evaluation Benchmark Discussion!
2024-02-20SymbolicAI / ExtensityAI Paper Overview (Part 1) - Key Philosophy Behind the Design - Symbols
2024-02-13Embeddings Walkthrough (Part 2): Context-Dependent Embeddings, Shifting Embedding Space
2024-02-06Embeddings Walkthrough (Part 1) - Bag of Words to word2vec to Transformer contextual embeddings
2024-01-29V* - Better than GPT-4V? Iterative Context Refining for Visual Question Answer!
2024-01-23AutoGen: A Multi-Agent Framework - Overview and Improvements
2024-01-09AppAgent: Using GPT-4V to Navigate a Smartphone!
2024-01-08Tutorial #13: StrictJSON, my first Python Package! - Get LLMs to output into a working JSON!
2023-12-20"Are you smarter than an LLM?" game speedrun
2023-12-08Is Gemini better than GPT4? Self-created benchmark - Fact Retrieval/Checking, Coding, Tool Use
2023-12-04Learning, Fast and Slow: 10 Years Plan - Memory Soup, Hier. Planning, Emotions, Knowledge Sharing
2023-12-01Tutorial #12: Use ChatGPT and off-the-shelf RAG on Terminal/Command Prompt/Shell - SymbolicAI
2023-11-20JARVIS-1: Multi-modal (Text + Image) Memory + Decision Making with LLMs in MineCraft!
2023-11-20Tutorial #11: Virtual Persona from Documents, Multi-Agent Chat, Text-to-Speech to hear your Personas
2023-11-14A Roadmap for AI: Past, Present and Future (Part 3) - Multi-Agent, Multiple Sampling and Filtering
2023-11-07Learning, Fast and Slow: My Landmark Idea for fast, adaptable agents (ICDL 2023 Best Paper Finalist)
2023-11-06A roadmap for AI: Past, Present and Future (Part 2): Fixed vs Flexible, Memory Soup vs Hierarchy
2023-11-03AI & Education: Education when AI tools are smarter than us - Discussion with Kuang Wen (Part 2)
2023-11-03AI & Education: RAG Question-Answer, Test Question Generator, Autograder by Kuang Wen! (Part 1)
2023-10-31A Roadmap for AI: Past, Present and Future (Part 1)
2023-10-28Tutorial #10: StrictJSON v2 (StrictText): Handle any output - quotation marks or backslash!