Skip to yearly menu bar Skip to main content


Contributed talks
in
Workshop: HiLD: High-dimensional Learning Dynamics Workshop

Contributed talks 2

Simon Du · Wei Huang · Yuandong Tian


Abstract:

Three (15 mins) contributed talks in this session.

  • Simon Du, Over-Parameterization Exponentially Slows Down Gradient Descent for Learning a Single Neuron

  • Wei Huang, Graph Neural Networks Provably Benefit from Structural Information: A Feature Learning Perspective

  • Yuandong Tian, Scan and Snap: Understanding Training Dynamics and Token Composition in 1-layer Transformer

Chat is not available.