Eshaan Nichani

6

Papers

5

Total Citations

Papers (6)

Learning Hierarchical Polynomials with Three-Layer Neural Networks

How Transformers Learn Causal Structure with Gradient Descent

Identifying good directions to escape the NTK regime and efficiently learn low-degree plus sparse polynomials

Smoothing the Landscape Boosts the Signal for SGD: Optimal Sample Complexity for Learning Single Index Models

Provable Guarantees for Nonlinear Feature Learning in Three-Layer Neural Networks

Fine-Tuning Language Models with Just Forward Passes