Anji Liu

OpenHA: A Series of Open-Source Hierarchical Agentic Models in Minecraft

Learning Tractable Distributions of Language Model Continuations

DeepHA: Scaling Action Chains Elicits Deep Hierarchical Agents

S-Chain: Structured Visual Chain-of-Thought for Medicine

Learning to Discretize Denoising Diffusion ODEs

Rao-Blackwell Gradient Estimators for Equivariant Denoising Diffusion

Plug-and-Play Context Feature Reuse for Efficient Masked Generation

Mitigating Reward Over-optimization in Direct Alignment Algorithms with Importance Sampling

Tractable Transformers for Flexible Conditional Generation

Discrete Copula Diffusion

GROOT-2: Weakly Supervised Multimodal Instruction Following Agents

Optimizing Inference-Time Reasoning in LLMs via Retrieval-Augmented Reflection

Training Open-ended Policies to follow Video-prompt Instructions with Reinforcement Learning