Jianwen Xie

Researcher@Lambda·美国·OpenReview

研究方向

unsupervised learning · representation learning · texture synthesis · energy-based model · 3d generation · generative model · video synthesis · latent variable model

Jianwen Xie

In-the-Flow Agentic System Optimization for Effective Planning and Tool Use

Principled RL for Diffusion LLMs Emerges from a Sequence-Level Perspective

VideoNSA: Native Sparse Attention Scales Video Understanding

EdiVal-Agent: An Object-Centric Framework for Automated, Fine-Grained Evaluation of Multi-Turn Editing

Align Your Structures: Generating Trajectories with Structure Pretraining for Molecular Dynamics

To Compress or Not? Pushing the Frontier of Lossless GenAI Model Weights Compression with Exponent Concentration

Randomized Antipodal Search Done Right for Data Pareto Improvement of LLM Unlearning

Mitigating Non-IID Drift in Zeroth-Order Federated LLM Fine-Tuning with Transferable Sparsity

``Noisier'’ Noise Contrastive Estimation is (Almost) Maximum Likelihood

InstructPro: Natural Language Guided Ligand-Binding Protein Design

AuxVLA: Auxiliary Task Learning and Multi-Modal Enhancement for Vision-Language-Action Models in Mobile Manipulation

Learning Materials Interatomic Potentials via Hybrid Invariant-Equivariant Architectures

FFT-Accelerated Auxiliary Variable MCMC for Fermionic Lattice Models: A Determinant-Free Approach with $O(N \log N)$ Complexity

SpatialReasoner: Towards Explicit and Generalizable 3D Spatial Reasoning

Tensor Decomposition Networks for Accelerating Machine Learning Force Field Computations

Latent Thought Models with Variational Bayes Inference-Time Computation