Eshaan Nichani

27 papers A* 10Journal 16Unranked 1
YearRankTypeTitle / Venue / Authors
2026 J jnl
Sharp Capacity Scaling of Spectral Optimizers in Learning Associative Memory.
CoRR
Juno Kim, Eshaan Nichani, Denny Wu, Alberto Bietti, Jason D. Lee
2025 J jnl
Emergence and scaling laws in SGD learning of shallow neural networks.
CoRR
Yunwei Ren, Eshaan Nichani, Denny Wu, Jason D. Lee
2025 A* conf
Learning Compositional Functions with Transformers from Easy-to-Hard Data.
COLT
Zixuan Wang, Eshaan Nichani, Alberto Bietti, Alex Damian, Daniel Hsu, Jason D. Lee, Denny Wu
2025 J jnl
Learning Compositional Functions with Transformers from Easy-to-Hard Data.
CoRR
Zixuan Wang, Eshaan Nichani, Alberto Bietti, Alex Damian, Daniel Hsu, Jason D. Lee, Denny Wu
2025 A* conf
Learning Hierarchical Polynomials of Multiple Nonlinear Features.
ICLR
Hengyu Fu, Zihao Wang, Eshaan Nichani, Jason D. Lee
2025 J jnl
On the Statistical Query Complexity of Learning Semiautomata: a Random Walk Approach.
CoRR
George Giapitzakis, Kimon Fountoulakis, Eshaan Nichani, Jason D. Lee
2025 J jnl
Quantitative Bounds for Length Generalization in Transformers.
CoRR
Zachary Izzo, Eshaan Nichani, Jason D. Lee
2025 A* conf
Understanding Factual Recall in Transformers via Associative Memories.
ICLR
Eshaan Nichani, Jason D. Lee, Alberto Bietti
2024 A* conf
How Transformers Learn Causal Structure with Gradient Descent.
ICML
Eshaan Nichani, Alex Damian, Jason D. Lee
2024 J jnl
How Transformers Learn Causal Structure with Gradient Descent.
CoRR
Eshaan Nichani, Alex Damian, Jason D. Lee
2024 J jnl
Learning Hierarchical Polynomials of Multiple Nonlinear Features with Three-Layer Networks.
CoRR
Hengyu Fu, Zihao Wang, Eshaan Nichani, Jason D. Lee
2024 A* conf
Learning Hierarchical Polynomials with Three-Layer Neural Networks.
ICLR
Zihao Wang, Eshaan Nichani, Jason D. Lee
2024 J jnl
Understanding Factual Recall in Transformers via Associative Memories.
CoRR
Eshaan Nichani, Jason D. Lee, Alberto Bietti
2023 A* conf
Fine-Tuning Language Models with Just Forward Passes.
NeurIPS
Sadhika Malladi, Tianyu Gao, Eshaan Nichani, Alex Damian, Jason D. Lee, Danqi Chen, Sanjeev Arora
2023 J jnl
Fine-Tuning Language Models with Just Forward Passes.
CoRR
Sadhika Malladi, Tianyu Gao, Eshaan Nichani, Alex Damian, Jason D. Lee, Danqi Chen, Sanjeev Arora
2023 J jnl
Learning Hierarchical Polynomials with Three-Layer Neural Networks.
CoRR
Zihao Wang, Eshaan Nichani, Jason D. Lee
2023 A* conf
Provable Guarantees for Nonlinear Feature Learning in Three-Layer Neural Networks.
NeurIPS
Eshaan Nichani, Alex Damian, Jason D. Lee
2023 J jnl
Provable Guarantees for Nonlinear Feature Learning in Three-Layer Neural Networks.
CoRR
Eshaan Nichani, Alex Damian, Jason D. Lee
2023 A* conf
Self-Stabilization: The Implicit Bias of Gradient Descent at the Edge of Stability.
ICLR
Alex Damian, Eshaan Nichani, Jason D. Lee
2023 A* conf
Smoothing the Landscape Boosts the Signal for SGD: Optimal Sample Complexity for Learning Single Index Models.
NeurIPS
Alex Damian, Eshaan Nichani, Rong Ge, Jason D. Lee
2023 J jnl
Smoothing the Landscape Boosts the Signal for SGD: Optimal Sample Complexity for Learning Single Index Models.
CoRR
Alex Damian, Eshaan Nichani, Rong Ge, Jason D. Lee
2022 — conf
Causal Structure Discovery between Clusters of Nodes Induced by Latent Factors.
CLeaR
Chandler Squires, Annie Yun, Eshaan Nichani, Raj Agrawal, Caroline Uhler
2022 A* conf
Identifying good directions to escape the NTK regime and efficiently learn low-degree plus sparse polynomials.
NeurIPS
Eshaan Nichani, Yu Bai, Jason D. Lee
2022 J jnl
Identifying good directions to escape the NTK regime and efficiently learn low-degree plus sparse polynomials.
CoRR
Eshaan Nichani, Yu Bai, Jason D. Lee
2022 J jnl
Self-Stabilization: The Implicit Bias of Gradient Descent at the Edge of Stability.
CoRR
Alex Damian, Eshaan Nichani, Jason D. Lee
2020 J jnl
Balancedness and Alignment are Unlikely in Linear Neural Networks.
CoRR
Adityanarayanan Radhakrishnan, Eshaan Nichani, Daniel Irving Bernstein, Caroline Uhler
2020 J jnl
Do Deeper Convolutional Networks Perform Better?
CoRR
Eshaan Nichani, Adityanarayanan Radhakrishnan, Caroline Uhler