↓Skip to main content

🏢 University of California, San Diego

Stable Minima Cannot Overfit in Univariate ReLU Networks: Generalization by Large Step Sizes
·2167 words·11 mins· loading · loading
AI Theory Generalization 🏢 University of California, San Diego
Deep ReLU networks trained with large, constant learning rates avoid overfitting in univariate regression due to minima stability, generalizing well even with noisy labels.
Probablistic Emulation of a Global Climate Model with Spherical DYffusion
·2553 words·12 mins· loading · loading
🏢 University of California, San Diego
Spherical DYffusion: a novel AI model generates accurate, physically consistent global climate ensemble simulations, surpassing existing methods in efficiency and skill.
MeshFormer : High-Quality Mesh Generation with 3D-Guided Reconstruction Model
·1982 words·10 mins· loading · loading
3D Vision 🏢 University of California, San Diego
MeshFormer: High-quality 3D mesh generation from sparse views in seconds, using transformers and 3D convolutions.
DiffTORI: Differentiable Trajectory Optimization for Deep Reinforcement and Imitation Learning
·2324 words·11 mins· loading · loading
Reinforcement Learning 🏢 University of California, San Diego
DiffTORI leverages differentiable trajectory optimization for superior deep reinforcement and imitation learning, outperforming prior state-of-the-art methods on high-dimensional robotic tasks.