PeopleofSuperintelligence

Researcher

Ming-Chang Chiu

Ming-Chang Chiu is a researcher whose work spans trustworthy AI, robust vision-language models, multimodal video generation, and AI agent systems. In 2020, he coauthored a machine-learning method for screenplay scoring; his 2022 papers examined robust representations and image-classification subgroups. He coauthored VideoPoet, which received ICML's 2024 Best Paper Award, and introduced AIDE in February 2025 to improve vision-language models with specialized domain-expert models. He defended his USC computer science dissertation in July 2025; as of October 2026, he is a Cornell postdoc and the founder and CEO of APAL.

Now at
Cornell University and APAL
Badges

Timeline

  1. 2020

    Predicting screenplay nominations

    In 2020, he coauthors a machine-learning method to score screenplay quality from linguistic cues and predict nominations for scripts at major film awards.

  2. February 2022

    Robust representations across shifts

    In February 2022, with Xuezhe Ma, he proposes an algorithm that combines adversarial training and group distribution robust optimization to improve robust representation learning.

  3. December 2022

    A study of image subgroups

    In December 2022, he and his coauthors study natural background colors as spurious features in image classification by grouping CIFAR-10 and CIFAR-100 test images by background color.

  4. 2024

    VideoPoet wins an ICML award

    In 2024, with Google Research colleagues, he coauthors VideoPoet, a multimodal language model that synthesizes video from images, video, text, and audio. VideoPoet receives ICML's 2024 Best Paper Award.

  5. February 2025

    AIDE for vision-language models

    In February 2025, he and his coauthors introduce AIDE, a framework that uses specialized domain-expert models to improve vision-language models.

  6. July 2025

    A USC dissertation defense

    In July 2025, he defends his USC computer science dissertation, titled “From Subgroup Discrepancy to Agentic Learning: A Data-Driven Approach to Building and Evaluating Robust and Contextually Aligned Vision-Language Models.”

  7. As of October 2026

    Founder and CEO of APAL

    He is a Cornell postdoc and founder and CEO of APAL, an AI assurance startup that builds trust infrastructure for autonomous agent-to-agent interactions. His research focuses on trustworthy AI, vision-language models, and agentic systems.

Career

  1. NowCornell UniversityPostdoc
  2. NowAPALFounder and CEO

Education

  1. 2025University of Southern CaliforniaPhD

From the biography.

  1. 11% of AI doctorate holders from 2010 to 2016 founded or led a company or lab by 10 years after graduationIncludes Cornell University. We estimate how often AI doctorate holders founded or led a company or lab by 10 years after graduation, how waiting times compare between founders and top jobs, and which universities trained the most.
  2. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Cornell University. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  3. 84% of people shaping AI with recorded undergraduate and doctoral schools changed institutions for their doctorateIncludes Cornell University. 84% of people shaping AI with recorded undergraduate and doctoral schools changed institutions for their doctorate. Across 2,116 people with both an undergraduate degree and a doctorate recorded, we trace which undergraduate schools feed which doctoral programs, how often students stay at the same school, and how paths differ by birth-country region.
  4. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Cornell University. 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.

Sources and credits

Photo: We have no published photo of Ming-Chang Chiu yet. Send us a free-licensed photo.

  1. Screenplay Quality Assessment: Can We Predict Who Gets Nominated? arxiv.org. Checked 21 October 2026.
  2. Learning Representations Robust to Group Shifts and Adversarial Examples arxiv.org. Checked 21 October 2026.
  3. Better May Not Be Fairer: A Study on Subgroup Discrepancy in Image Classification arxiv.org. Checked 21 October 2026.
  4. VideoPoet – Google Research sites.research.google. Checked 21 October 2026.
  5. ICML 2024 Awards icml.cc. Checked 21 October 2026.
  6. Ming-Chang Chiu | ETHDenver ethdenver.com. Checked 21 October 2026.
  7. AIDE: Agentically Improve Visual Language Model with Domain Experts arxiv.org. Checked 21 October 2026.
  8. PhD Dissertation Defense - Ming-Chang Chiu viterbi.usc.edu. Checked 21 October 2026.