Ryo Karakida
11 papers · 2019–2025 · 4 conferences · across top CS/AI conferences
Achievements
Jump to papers ↓+4 more ↓ Show less ↑
π Interdisciplinary Bridge π§ Keyword Pioneer π Academic Marathon (6) π Cross-Pollinator (12) πΊοΈ Taxonomy Completionist (11)
π
Conference Polyglot
(4)
π
Keyword Champion
(4)
π
Century Club
(11)
π₯
Unstoppable
(7)
Conferences
AISTATS (3)
ICLR (3)
ICML (3)
NIPS (2)
Top co-authors
Keywords
fisher information matrix
(4)
neural network optimization
(3)
deep neural network
(2)
convergence analysis
(1)
deep learning
(1)
loss landscape
(1)
natural gradient
(1)
fisher information
(1)
finite difference
(1)
natural gradient descent
(1)
generalization ability
(1)
learning rate
(1)
implicit bia
(1)
flat minima
(1)
generalization performance
(1)
layer normalization
(1)
diagonal linear network
(1)
dynamical isometry
(1)
spectral distribution
(1)
free probability theory
(1)
Papers
Local Loss Optimization in the Infinite Width: Stable Parameterization of Predictive Coding Networks and Target Propagation
ICLR 2025
On the Parameterization of Second-Order Optimization Effective towards the Infinite Width
ICLR 2024
Understanding MLP-Mixer as a wide and sparse MLP
ICML 2024
Self-attention Networks Localize When QK-eigenspectrum Concentrates
ICML 2024
Understanding Gradient Regularization in Deep Learning: Efficient Finite-Difference Computation and Implicit Bias
ICML 2023
Learning curves for continual learning in neural networks: Self-knowledge transfer and forgetting
ICLR 2022
The Spectrum of Fisher Information of Deep Networks Achieving Dynamical Isometry
AISTATS 2021
Understanding Approximate Fisher Information for Fast Convergence of Natural Gradient Descent in Wide Neural Networks
NIPS 2020
The Normalization Method for Alleviating Pathological Sharpness in Wide Neural Networks
NIPS 2019
Universal Statistics of Fisher Information in Deep Neural Networks: Mean Field Approach
AISTATS 2019
Fisher Information and Natural Gradient Learning in Random Deep Networks
AISTATS 2019