Yi Ma

Pensieve: Self-supervised Novel View Synthesis via Implicit and Explicit Reconstruction

Implicit Neural Representation Generation with Hypernetworks

Flow Policy Gradients for Legged Robots

CUPID: Pose-Grounded Generative 3D Reconstruction from a Single Image

Pointer-CAD: Unifying B-Rep and Command Sequences via Pointer-based Edges & Faces Selection

CodePlot-CoT: Mathematical Visual Reasoning by Thinking with Code-Driven Images

Generative Simulation for Dexterous Hands

Token Statistics Transformer: Linear-Time Attention via Variational Rate Reduction

Cameras as Relative Positional Encoding

On the Edge of Memorization in Diffusion Models

Attention-Only Transformers via Unrolled Subspace Denoising

Simplifying DINO via Coding Rate Regularization

Attention-Only Transformers via Unrolled Subspace Denoising

SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training

On Evaluation of Generative Robotic Simulations

Improving Neuron-level Interpretability with White-box Language Models

Diffusion Models Learn Low-Dimensional Distributions via Subspace Clustering

Ctrl123: Consistent Novel View Synthesis via Closed-Loop Transcription