Sotabase

Career

· Miller Postdoctoral Fellow, UC Berkeley2025–
· Intern, NTT Research2021–2021
· PhD in Computational and Mathematical Engineering, Stanford University2020–2025
· Intern, Broad Institute of MIT and Harvard2018–2018
· Course Assistant for CS230: Deep Learning, Stanford University2018–2019
· MS in Computational and Mathematical Engineering, Stanford University2017–2020
· Undergraduate Teaching Assistant in Computer Science, Brown University2016–2017
· Member, Stanford-Brown iGEM Team, NASA Ames Research Center2015–2015
· A.B. in Computational Biology, Brown University2013–2017
· Sc.B. in Applied Mathematics, Brown University2013–2017
· PhD Student, UC Berkeley Miller Fellow

Publications (29)

Pruning neural networks without any data by iteratively conserving synaptic flow
Neural Information Processing Systems · 2020
769
cited
Loss Landscapes of Regularized Linear Autoencoders
International Conference on Machine Learning · 2019
99
cited
Neural Mechanics: Symmetry and Broken Conservation Laws in Deep Learning Dynamics
International Conference on Learning Representations · 2020
92
cited
Caffeine promotes ethanol drinking in rats. Examination using a limited-access free choice paradigm.
Alcohol · 2000
74
cited
Noether's Learning Dynamics: Role of Symmetry Breaking in Neural Networks
Neural Information Processing Systems · 2021
40
cited
Stochastic collapse: how gradient noise attracts SGD dynamics towards simpler subnetworks
Neural Information Processing Systems · 2023
39
cited
Beyond the Quadratic Approximation: The Multiscale Structure of Neural Network Loss Landscapes
Journal of Machine Learning · 2022
37
cited
Two Routes to Scalable Credit Assignment without Weight Symmetry
International Conference on Machine Learning · 2020
36
cited
Nicotine and Ethanol Interaction on Conditioned Taste Aversions Induced by Both Drugs
Pharmacology, Biochemistry and Behavior · 1999
35
cited
Get rich quick: exact solutions reveal how unbalanced initializations promote rapid feature learning
Neural Information Processing Systems · 2024
28
cited
The Asymmetric Maximum Margin Bias of Quasi-Homogeneous Neural Networks
International Conference on Learning Representations · 2022
23
cited
The Limiting Dynamics of SGD: Modified Loss, Phase-Space Oscillations, and Anomalous Diffusion
Neural Computation · 2021
19
cited
Preexposure Effects of Nicotine and Acetaldehyde on Conditioned Taste Aversion Induced by Both Drugs
Pharmacology, Biochemistry and Behavior · 2000
16
cited
Differences in locomotor response to an inescapable novel environment predict sensitivity to aversive effects of amphetamine
Behavioural Pharmacology · 2001
14
cited
Cocaine and Ethanol Interaction in the Conditioned Taste Aversion Paradigm
Physiology and Behavior · 1999
13
cited
Caffeine promotes an ethanol-induced conditioned taste aversion: a dose-dependent interaction.
Experimental and clinical psychopharmacology · 2001
10
cited
Rethinking the limiting dynamics of SGD: modified loss, phase space oscillations, and anomalous diffusion
arXiv.org · 2021
9
cited
Noether's Learning Dynamics: The Role of Kinetic Symmetry Breaking in Deep Learning
arXiv.org · 2021
7
cited
Augmentation of corticosterone release by means of a caffeine-ethanol interaction in rats.
Alcohol · 2000
4
cited
Caffeine, nicotine and mecamylamine share stimulus properties in the preexposure conditioned taste aversion procedure
Psychopharmacology · 2001
4
cited
Sotabase
Daniel Kunin | Researcher Profile | Sotabase | Sotabase