Machine Learning Street Talk (MLST) Harri Valpola: System 2 AI and Planning in Model-Based Reinforcement Learning
May 25, 2020
Harri Valpola, the CEO and Founder of Curious AI, specializes in optimizing industrial processes through advanced AI. In this discussion, he dives into the fascinating world of System 1 and System 2 thinking in AI, illustrating the balance between instinctive and reflective reasoning. Valpola shares insights from his recent research on model-based reinforcement learning, emphasizing the challenges of real-world applications like water treatment. He also highlights innovative approaches using denoising autoencoders to improve planning in uncertain environments.
AI Snips
Chapters
Books
Transcript
Episode notes
System 1 vs. System 2 Thinking
- System 1 thinking is unconscious and intuitive, relying on crystallized knowledge gained from experience, much like current AI.
- System 2 thinking, however, involves internal simulations, enabling imagination, planning, and handling of novel situations.
Building an AI Business
- Building a successful AI company requires finding problems where advanced machine learning is the best solution.
- Often, simpler methods like linear regression suffice, making it crucial to identify the right niche for cutting-edge AI.
Humans and Monkeys Learning
- In a card classification experiment, humans and monkeys gradually learned to classify cards.
- However, humans could sometimes deduce the underlying rule, leading to an immediate jump to 100% accuracy, demonstrating System 2 thinking.
