Daniel Kunin
9 papers · 2019–2025 · 3 conferences · across top CS/AI conferences
Achievements
Jump to papers ↓+3 more ↓ Show less ↑
🐣 Hot Topic Early Bird 🧭 Keyword Pioneer 🌍 Conference Polyglot (3) 🏃 Academic Marathon (6) 🐝 Cross-Pollinator (13)
🌉
Interdisciplinary Bridge
🗺️
Taxonomy Completionist
(19)
📈
Trend Setter
Conferences
NIPS (4)
ICLR (3)
ICML (2)
Top co-authors
Keywords
learning dynamics
(2)
gradient descent
(2)
neural network
(2)
lottery ticket hypothesis
(1)
stochastic gradient descent
(1)
feature learning
(1)
variational inference
(1)
principal component analysis
(1)
local learning
(1)
sample complexity
(1)
probabilistic pca
(1)
sparse optimization
(1)
learning rate
(1)
convolutional neural network
(1)
linear autoencoder
(1)
deep neural network
(1)
implicit bia
(1)
linear network
(1)
gradient noise
(1)
credit assignment
(1)
Papers
From Lazy to Rich: Exact Learning Dynamics in Deep Linear Networks
ICLR 2025
Get rich quick: exact solutions reveal how unbalanced initializations promote rapid feature learning
NIPS 2024
Stochastic Collapse: How Gradient Noise Attracts SGD Dynamics Towards Simpler Subnetworks
NIPS 2023
The Asymmetric Maximum Margin Bias of Quasi-Homogeneous Neural Networks
ICLR 2023
Noether’s Learning Dynamics: Role of Symmetry Breaking in Neural Networks
NIPS 2021
Neural Mechanics: Symmetry and Broken Conservation Laws in Deep Learning Dynamics
ICLR 2021
Two Routes to Scalable Credit Assignment without Weight Symmetry
ICML 2020
Pruning neural networks without any data by iteratively conserving synaptic flow
NIPS 2020
Loss Landscapes of Regularized Linear Autoencoders
ICML 2019