Lechao Xiao
18 papers · 2018–2025 · 4 conferences · across top CS/AI conferences
Achievements
Jump to papers ↓+9 more ↓ Show less ↑
π Renaissance Researcher (5) π Interdisciplinary Bridge π Academic Marathon (7) π Conference Polyglot (4) πΊοΈ Taxonomy Completionist (20)
π§
Keyword Pioneer
π£
Hot Topic Early Bird
π
Conference Polyglot
(4)
π
Triple Crown
π€
Dynamic Duo
(14)
β‘
Prolific Year
(5)
π
Century Club
(18)
ποΈ
Keyword Collector
(52)
π₯
Unstoppable
(5)
Conferences
NIPS (7)
ICLR (5)
ICML (5)
COLT (1)
Top co-authors
Keywords
neural tangent kernel
(4)
convolutional neural network
(3)
neural network
(3)
generalization error
(2)
infinite width limit
(2)
linear model
(2)
kernel methods
(2)
convolutional network
(2)
statistical learning theory
(1)
gaussian process
(1)
mean field theory
(1)
gradient descent
(1)
image classification
(1)
spectral norm
(1)
asymptotic analysis
(1)
theoretical analysis
(1)
double descent
(1)
generalization bound
(1)
learning algorithm
(1)
eigenvalue analysis
(1)
Papers
Scaling Collapse Reveals Universal Dynamics in Compute-Optimally Trained Neural Networks
ICML 2025
Small-scale proxies for large-scale Transformer training instabilities
ICLR 2024
4+3 Phases of Compute-Optimal Neural Scaling Laws
NIPS 2024
Scaling Exponents Across Parameterizations and Optimizers
ICML 2024
Fast Neural Kernel Embeddings for General Activations
NIPS 2022
Synergy and Symmetry in Deep Learning: Interactions between the Data, Model, and Inference Algorithm
ICML 2022
Precise Learning Curves and Higher-Order Scalings for Dot-product Kernel Regression
NIPS 2022
Eigenspace Restructuring: A Principle of Space and Frequency in Neural Networks
COLT 2022
Dataset Distillation with Infinitely Wide Convolutional Networks
NIPS 2021
Exploring the Uncertainty Properties of Neural Networksβ Implicit Priors in the Infinite-Width Limit
ICLR 2021
Neural Tangents: Fast and Easy Infinite Neural Networks in Python
ICLR 2020
Finite Versus Infinite Neural Networks: an Empirical Study
NIPS 2020
The Surprising Simplicity of the Early-Time Learning Dynamics of Neural Networks
NIPS 2020
Provable Benefit of Orthogonal Initialization in Optimizing Deep Linear Networks
ICLR 2020
Disentangling Trainability and Generalization in Deep Neural Networks
ICML 2020
Wide Neural Networks of Any Depth Evolve as Linear Models Under Gradient Descent
NIPS 2019
Bayesian Deep Convolutional Networks with Many Channels are Gaussian Processes
ICLR 2019
Dynamical Isometry and a Mean Field Theory of CNNs: How to Train 10,000-Layer Vanilla Convolutional Neural Networks
ICML 2018