PeopleofSuperintelligence

Researcher

Teng Xiao

Teng Xiao is a researcher whose work spans graph learning, language-model alignment, diffusion models and reinforcement learning. His 2022 graph-learning paper earned a NeurIPS Spotlight, and in 2024 he coauthored Cal-DPO for language-model alignment. In 2025, he coauthored DSPO for diffusion-model alignment, completed a PhD in informatics at Penn State under Vasant Honavar, and received an honorable mention in the university's AI dissertation awards. He joined the Allen Institute for AI in 2025 and contributed to reinforcement-learning post-training for Olmo 3, a fully open family of language models at 7B and 32B scales. As of October 2026, he is a postdoctoral researcher at the University of Washington's H2Lab and works at AI2 and the university with Professors Noah Smith and Hanna Hajishirzi.

Works
53
Citations
515
h-index
11

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2022

    Graph self-supervision

    In 2022, he coauthors Decoupled Self-supervised Learning for Graphs, a framework for self-supervised node representation learning on graphs. NeurIPS selects the paper as a Spotlight.

  2. 2023

    GraphACL for graph learning

    In 2023, he coauthors Simple and Asymmetric Graph Contrastive Learning without Augmentations, which introduces GraphACL for representation learning on homophilic and heterophilic graphs. It uses an asymmetric view of neighboring nodes.

  3. 2024

    Cal-DPO for language models

    In 2024, he coauthors Cal-DPO to align large language models with human preference data. The method calibrates implicit reward values to improve alignment.

  4. 2025

    DSPO for diffusion alignment

    With Huaisheng Zhu and Vasant G. Honavar, he coauthors DSPO in 2025 to align text-to-image diffusion models with human preferences. The method uses score matching, and ICLR selects the paper as an Oral.

  5. 2025

    PhD and AI dissertation honor

    In 2025, he completes a PhD in informatics at Penn State under adviser Vasant Honavar. Penn State gives him an honorable mention in its AI dissertation awards for work on improving AI systems and their interactions with users.

  6. Since 2025

    Joining AI2 as a researcher

    In 2025, he joins the Allen Institute for AI as a researcher.

  7. 2025

    Olmo 3 post-training

    In 2025, he is among Olmo 3’s authors and serves as a core contributor to reinforcement-learning post-training. The fully open language-model family includes 7B and 32B parameter scales.

  8. As of October 2026

    Postdoc at UW’s H2Lab

    As of October 2026, he is a postdoctoral researcher at the University of Washington’s H2Lab. He works at AI2 and the University of Washington with Professors Noah A. Smith and Hanna Hajishirzi.

Career

  1. Since 2025Allen Institute for AI (Ai2)Researcher
  2. Since 2025University of WashingtonPostdoctoral Researcher

Education

  1. 2025Pennsylvania State UniversityPhD

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

  1. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Allen Institute for AI (Ai2). We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  2. The median h-index is 48 at universities and 14 at frontier AI labsIncludes University of Washington. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  3. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Allen Institute for AI (Ai2). Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.
  4. Of 943 people shaping AI who worked at Google, 446 work there nowIncludes University of Washington. Our estimates of where the 943 people shaping AI who worked at Google and Google DeepMind work now: 446 at Google and 452 across 278 other organizations.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Teng Xiao yet. Send us a free-licensed photo.

  1. Decoupled Self-supervised Learning for Graphs proceedings.neurips.cc. Checked 22 October 2026.
  2. Teng Xiao tengxiao1.github.io. Checked 22 October 2026.
  3. [2310.18884] Simple and Asymmetric Graph Contrastive Learning without Augmentations arxiv.org. Checked 22 October 2026.
  4. Cal-DPO: Calibrated Direct Preference Optimization for Language Model Alignment papers.nips.cc. Checked 22 October 2026.
  5. DSPO: DIRECT SCORE PREFERENCE OPTIMIZATION FOR DIFFUSION MODEL ALIGNMENT - Penn State pure.psu.edu. Checked 22 October 2026.
  6. Doctoral students honored for AI dissertations - AI at Penn State ai.psu.edu. Checked 22 October 2026.
  7. [2512.13961] Olmo 3 arxiv.org. Checked 22 October 2026.
  8. People h2lab.cs.washington.edu. Checked 22 October 2026.