Daniel Kunin

Publication Activity (10 Years)

Years Active: 2019-2024
Publications (10 Years): 17

Top Topics

Gradient Direction

Top Venues

Publications

Daniel Kunin, Allan Raventós, Clémentine Dominé, Feng Chen, David Klindt, Andrew M. Saxe, Surya Ganguli
Get rich quick: exact solutions reveal how unbalanced initializations promote rapid feature learning. CoRR (2024)
Daniel Kunin, Javier Sagastuy-Breña, Lauren Gillespie, Eshed Margalit, Hidenori Tanaka, Surya Ganguli, Daniel L. K. Yamins
The Limiting Dynamics of SGD: Modified Loss, Phase-Space Oscillations, and Anomalous Diffusion. Neural Comput. 36 (1) (2024)
Feng Chen, Daniel Kunin, Atsushi Yamamura, Surya Ganguli
Stochastic Collapse: How Gradient Noise Attracts SGD Dynamics Towards Simpler Subnetworks. NeurIPS (2023)
Daniel Kunin, Atsushi Yamamura, Chao Ma, Surya Ganguli
The Asymmetric Maximum Margin Bias of Quasi-Homogeneous Neural Networks. ICLR (2023)
Feng Chen, Daniel Kunin, Atsushi Yamamura, Surya Ganguli
Stochastic Collapse: How Gradient Noise Attracts SGD Dynamics Towards Simpler Subnetworks. CoRR (2023)
Daniel Kunin, Atsushi Yamamura, Chao Ma, Surya Ganguli
The Asymmetric Maximum Margin Bias of Quasi-Homogeneous Neural Networks. CoRR (2022)
Hidenori Tanaka, Daniel Kunin
Noether's Learning Dynamics: Role of Symmetry Breaking in Neural Networks. NeurIPS (2021)
Daniel Kunin, Javier Sagastuy-Breña, Surya Ganguli, Daniel L. Yamins, Hidenori Tanaka
Neural Mechanics: Symmetry and Broken Conservation Laws in Deep Learning Dynamics. ICLR (2021)
Daniel Kunin, Javier Sagastuy-Breña, Lauren Gillespie, Eshed Margalit, Hidenori Tanaka, Surya Ganguli, Daniel L. K. Yamins
Rethinking the limiting dynamics of SGD: modified loss, phase space oscillations, and anomalous diffusion. CoRR (2021)
Hidenori Tanaka, Daniel Kunin
Noether's Learning Dynamics: The Role of Kinetic Symmetry Breaking in Deep Learning. CoRR (2021)
Daniel Kunin, Aran Nayebi, Javier Sagastuy-Breña, Surya Ganguli, Jonathan M. Bloom, Daniel L. K. Yamins
Two Routes to Scalable Credit Assignment without Weight Symmetry. CoRR (2020)
Daniel Kunin, Javier Sagastuy-Breña, Surya Ganguli, Daniel L. K. Yamins, Hidenori Tanaka
Neural Mechanics: Symmetry and Broken Conservation Laws in Deep Learning Dynamics. CoRR (2020)
Hidenori Tanaka, Daniel Kunin, Daniel L. Yamins, Surya Ganguli
Pruning neural networks without any data by iteratively conserving synaptic flow. NeurIPS (2020)
Daniel Kunin, Aran Nayebi, Javier Sagastuy-Breña, Surya Ganguli, Jonathan M. Bloom, Daniel Yamins
Two Routes to Scalable Credit Assignment without Weight Symmetry. ICML (2020)
Hidenori Tanaka, Daniel Kunin, Daniel L. K. Yamins, Surya Ganguli
Pruning neural networks without any data by iteratively conserving synaptic flow. CoRR (2020)
Daniel Kunin, Jonathan M. Bloom, Aleksandrina Goeva, Cotton Seed
Loss Landscapes of Regularized Linear Autoencoders. ICML (2019)
Daniel Kunin, Jonathan M. Bloom, Aleksandrina Goeva, Cotton Seed
Loss Landscapes of Regularized Linear Autoencoders. CoRR (2019)