Longxuan Yu (Ted)

About Me

Hi! I am Longxuan (Ted) Yu, a second-year Ph.D. student (2025.09 – ) in the Department of Computer Science and Engineering at the University of California, Riverside (UCR). My advisor is Prof. Greg Ver Steeg.

Before joining UCR, I received my M.S. in Data Science from the University of California, San Diego (UCSD), where I worked with Prof. Haojian Jin on trustworthy AI and generative models. I obtained my B.Eng. in Computer Science from Hangzhou Dianzi University (HDU).

My current research interests lie in Diffusion Models, Representation Learning, and Information-theoretic Methods. I am particularly interested in the theoretical foundations of generative modeling, mutual information estimation, and how representation learning can support generalization across modalities and domains.

If you want to know more about my research or potential collaborations, feel free to email me at ylong030@ucr.edu.

Publications

UAI 2026

MMG: Mutual Information Estimation via the MMSE Gap in Diffusion

Longxuan Yu, Xing Shi, Xianghao Kong, Tong Jia, Greg Ver Steeg

Conference on Uncertainty in Artificial Intelligence (UAI), 2026. Poster.

arXiv Paper Code Package
ICML 2026

Is Your Diffusion Sampler Actually Correct? A Sampler-Centric Evaluation of Discrete Diffusion Language Models

Luhan Tang, Longxuan Yu, Shaorong Zhang, Greg Ver Steeg

International Conference on Machine Learning (ICML), 2026.

arXiv Paper Project Code
NeurIPS 2026

Discrete Stochastic Localization for Non-autoregressive Generation

Yunshu Wu, Jiayi Cheng, Longxuan Yu, Partha Thakuria, Rob Brekelmans, Evangelos E. Papalexakis, Greg Ver Steeg

Advances in Neural Information Processing Systems (NeurIPS), 2026. Poster.

arXiv Blog
NeurIPS 2026

Harbor Adapters and Harbor-Index: Infrastructure and a Curated Meta-Dataset for Large-Scale Agentic Evaluation

Lin Shi, Haowei Lin, Zixuan Zhu, Xiaoyue Zhou, Xiang Li, ..., Longxuan Yu, et al.

NeurIPS 2026 Evaluations and Datasets Track. Poster.

arXiv Project Code
SIGCHI 2026

PrivacyAkinator: Articulating Key Privacy Design Decisions by Answering LLM-Generated Multiple-choice Questions

Qiyu Li, Yuen Sum Wong, Yuen Kei Wong, Longxuan Yu, Haojian Jin

ACM CHI, 2026.

arXiv Paper Code
EMNLP 2026

DSL-LLaDA: Scaling Continuous Denoising to 8B Masked Diffusion LMs

Longxuan Yu, Yunshu Wu, Yu Fu, Siheng Xiong, Rob Brekelmans, Hui Liu, Yue Dong, Greg Ver Steeg

Findings of the Association for Computational Linguistics: EMNLP 2026.

arXiv Demo Code
arXiv 2026

Revise, Don't Freeze: Sampler-Matched Training for Self-Correcting Masked Diffusion Language Models

Longxuan Yu, Shaorong Zhang, Yu Fu, Hui Liu, Yue Dong, Greg Ver Steeg

arXiv preprint arXiv:2606.01026, 2026.

arXiv
arXiv 2026

Reducing the Safety Tax in LLM Safety Alignment with On-Policy Self-Distillation

Yu Fu, Longxuan Yu, Haz Sameen Shahgir, Zhipeng Wei, Hui Liu, N. Benjamin Erichson, Yue Dong

arXiv preprint arXiv:2605.15239, 2026.

arXiv
arXiv 2026

Generation Order and Parallel Decoding in Masked Diffusion Models: An Information-Theoretic Perspective

S Zhang, Longxuan Yu, R Brekelmans, L Tang, S Asif, GV Steeg

arXiv preprint arXiv:2602.00286, 2026.

arXiv
arXiv 2026

Thinking Out of Order: When Output Order Stops Reflecting Reasoning Order in Diffusion Language Models

Longxuan Yu, Y Fu, S Zhang, H Liu, GV Steeg, Y Dong

arXiv preprint arXiv:2601.22035, 2026.

arXiv Code
NAACL 2025

CausalEval: Towards Better Causal Reasoning in Language Models

Longxuan Yu, Delin Chen, Siheng Xiong, Qingyang Wu, Qingzhen Liu, Dawei Li, Zhikai Chen, Xiaoze Liu, Liangming Pan

NAACL, 2025.

arXiv Resources
ACM CCS 2024

Moderator: Moderating Text-to-Image Diffusion Models through Fine-grained Context-based Policies

Peiran Wang, Qiyu Li, Longxuan Yu, Ziyao Wang, Ang Li, Haojian Jin

ACM CCS, 2024. Distinguished Paper Award

arXiv Code

Academic Experience

2025.09 – present, Ph.D. Student, Department of Computer Science and Engineering, UCR
Advisor: Prof. Greg Ver Steeg Research Area: Diffusion models, Representation learning, Information-theoretic methods
2023.09 – 2025.06, Research Assistant, Halıcıoğlu Data Science Institute (HDSI), UCSD
Advisor: Prof. Haojian Jin
Project: Moderator: Moderating Text-to-Image Diffusion Models through Fine-grained Context-based Policies
Research Area: Diffusion-based generative modeling, Human-centered trustworthy AI

Industry & Research Experience

2026.06 – present, Applied Scientist Intern, Amazon Web Services (AWS), New York
2025.03 – 2025.08, Research Assistant, MiraclePlus, Beijing
2023.06 – 2023.09, Software Engineer Intern / OSPP, Xinference (formerly Xorbits)

Open-Source & Security Contributions

Linux Kernel — CVE-2026-46153 (VLAN / net/8021q): Authored the upstream fix 8021q: delete cleared egress QoS mappings, co-developed with Yuan Tan. CVE advisory
Linux Kernel — CVE-2026-52933 (io_uring/poll): Authored the upstream fix io_uring/poll: fix signed comparison in io_poll_get_ownership(), co-developed with Yuan Tan. CVE advisory
Harbor-Index 1.0 — Contributor: Listed as an official Harbor-Index contributor. Built the FeatureBench adapter and contributed benchmark dataset conversion, validation, and fixes; four FeatureBench tasks are included in the final 82-task index. Co-author of Harbor Adapters and Harbor-Index: Infrastructure and a Curated Meta-Dataset for Large-Scale Agentic Evaluation, accepted to the NeurIPS 2026 Evaluations and Datasets Track (poster).
Xorbits / OSPP 2023 — Slurm Deployment: Initiated and developed Slurm deployment support for running Xorbits on HPC clusters, including sbatch script generation, job lifecycle management, testing infrastructure, and documentation. Released in Xorbits v0.7.1. Pull Request

Education Experience

2025.09 – present, Ph.D., Computer Science, UCR, USA
2023.09 – 2025.06, M.S., Data Science, UCSD, USA
2019.09 – 2023.06, B.Eng., Computer Science, Hangzhou Dianzi University (HDU), China

Honors and Awards

2024, ACM CCS Distinguished Paper Award (Moderator: Moderating Text-to-Image Diffusion Models through Fine-grained Context-based Policies)
2023, Outstanding Graduate, Hangzhou Dianzi University
2022, National Scholarship