Machine Learning Street Talk (MLST)

How Machines Learn to Ignore the Noise (Kevin Ellis + Zenna Tavares)

239 snips
Apr 8, 2025
Prof. Kevin Ellis, an AI and cognitive science expert at Cornell University, and Dr. Zenna Tavares, co-founder of BASIS, explore how AI can learn like humans. They discuss how machines can generate knowledge from minimal data through exploration and experimentation. The duo highlights the importance of compositionality, building complex ideas from simple ones, and the need for AI to grasp abstraction without getting lost in details. By blending different learning methods, they envision smarter AI that can tackle real-world challenges more intuitively.
Ask episode
AI Snips
Chapters
Books
Transcript
Episode notes
INSIGHT

Compositionality's Double-Edged Sword

  • Compositionality in AI allows combining small pieces of knowledge to solve complex problems.
  • However, it can lead to a combinatorial explosion of possibilities, making it hard to find relevant concepts.
INSIGHT

Compositionality in Programming Languages

  • Programming languages offer a good example of compositional structure in practice.
  • Unlike natural language, programming languages are strictly compositional, which simplifies their use but may limit expressiveness.
INSIGHT

Transduction vs. Induction

  • Some problems are better solved by intuition (transduction) than by explicit reasoning (induction).
  • LLMs, like humans, show varied performance across problem types with this split.
Get the Snipd Podcast app to discover more snips from this episode
Get the app