Papers
- Addressable Memory for Video World Models
- Variance Reduction for Expectations with Diffusion Teachers
- NVIDIA OmniDreams: Real-Time Generative World Model for Closed-Loop Autonomous Vehicle Simulation
- Motion Attribution for Video Generation
- Score Distillation Sampling for Audio: Source Separation, Synthesis, and Beyond
- LLaMA-Mesh: Unifying 3D Mesh Generation with Language Models
- Multi-Student Diffusion Distillation for Better One-Step Generators
- Improving Hyperparameter Optimization with Checkpointed Model Weights
- PhD Thesis: Scalable Nested Optimization for Deep Learning
- LATTE3D: Large-Scale Amortized Text-to-Enhanced3D Synthesis
- Training Data Attribution via Approximate Unrolled Differentiation
- Graph Metanetworks for Processing Diverse Neural Architectures
- Using Large Language Models for Hyperparameter Optimization
- ATT3D: Amortized Text-to-3D Object Synthesis
- Task Selection for AutoML System Evaluation
- On Implicit Bias in Overparameterized Bilevel Optimization
- Lyapunov Exponents for Diversity in Differentiable Games
- Complex Momentum for Optimization in Games
- Input Convex Gradient Networks
- Meta-Learning to Improve Pre-Training
- Using Bifurcations for Diversity in Differentiable Games
- Optimizing Millions of Hyperparameters by Implicit Differentiation
- JacNet: Learning Functions with Structured Jacobians
- Self-Tuning Networks: Bilevel Optimization of Hyperparameters using Structured Best-Response Functions
- Understanding Neural Architecture Search Techniques
- Stochastic Hyperparameter Optimization Through Hypernetworks