Physics for Future From Physics to Computer Science: Symmetry in Neural Networks with Prof. Tess Smidt
Apr 2, 2025
In this captivating chat, Professor Tess Smidt, an MIT Assistant Professor with a unique blend of physics and architecture, shares her insights on symmetry in neural networks. She explores how integrating symmetry enhances learning and scientific modeling. The discussion touches on the balance between quantum computing and neural networks, along with the intricacies of AI, machine learning, and deep learning. Tess also looks ahead to future innovations in geometric deep learning, promising groundbreaking applications in areas like self-driving cars and drug design.
AI Snips
Chapters
Transcript
Episode notes
Metacognition Passion
- Tess Smidt was fascinated with understanding how people think and explored this through a podcast and reading psychology and business books.
- Though not her profession, she remains passionate about metacognition and educating herself on it.
Field Hopping Motivated by Frustrations
- Tess shifted fields when frustrated by the lack of neural networks handling 3D geometry needed in her physics research.
- Her approach shows how interdisciplinary needs can drive career evolution and new innovations.
The Jargon Translation Challenge
- Technical jargon helps communication but also risks obscuring true meaning within interdisciplinary groups.
- Actively translating terms across fields enhances understanding and enriches research collaboration.