PeopleofSuperintelligence

Researcher

Jiajun Wu

Jiajun Wu is an assistant professor of computer science at Stanford, with a courtesy appointment in psychology. He works on physical scene understanding, which means building machines that see, reason about, and interact with the physical world. He co-wrote MarrNet in 2017 and the Neuro-Symbolic Concept Learner in 2019. He won the 2019 AAAI/ACM doctoral dissertation prize, and Samsung named him its 2020 AI Researcher of the Year. IJCAI recognized his work across computer vision, generative AI and robotics with its 2026 Computers and Thought Award, and before Stanford he was a visiting faculty researcher at Google Research.

Timeline

  1. 2017

    MarrNet for 3D shapes

    In 2017 he shares equal first authorship of MarrNet, a paper on 3D shape reconstruction via 2.5D sketches.

  2. 2019

    A neuro-symbolic concept learner

    In 2019 he co-authors the Neuro-Symbolic Concept Learner, a model that learns visual concepts and words from images and paired questions and answers. The paper appears at ICLR 2019.

  3. 2019–2020

    AAAI/ACM and MIT dissertation prizes

    He wins the 2019 AAAI/ACM doctoral dissertation prize. MIT gives him the George M. Sprowls PhD Thesis Award in 2020.

  4. 2020

    Samsung AI Researcher of the Year

    Samsung names him its 2020 AI Researcher of the Year.

  5. February 2026

    Physical scene understanding

    In an AIhub interview published in February 2026, he describes his research as physical scene understanding: building machines that see, reason about, and interact with the physical world.

  6. 2026

    The IJCAI Computers and Thought Award

    In 2026 IJCAI names him the winner of its Computers and Thought Award. IJCAI recognizes his work across computer vision, generative AI, and robotics.

  7. As of October 2026

    A professorship at Stanford

    He is an assistant professor of computer science and, by courtesy, of psychology at Stanford, working on computer vision, machine learning and computational cognitive science. He comes to Stanford from a visiting faculty post at Google Research.

Sources and credits

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

  1. [1711.03129] MarrNet: 3D Shape Reconstruction via 2.5D Sketches arxiv.org. Checked 14 October 2026.
  2. [1904.12584] The Neuro-Symbolic Concept Learner: Interpreting Scenes, Words, and Sentences From Natural Supervision arxiv.org. Checked 14 October 2026.
  3. The neuro-symbolic concept learner: Interpreting scenes, words, and sentences from natural supervision dspace.mit.edu. Checked 14 October 2026.
  4. Learning to see the physical world: an interview with Jiajun Wu - ΑΙhub aihub.org. Checked 14 October 2026.
  5. Jiajun Wu's Profile | Stanford Profiles profiles.stanford.edu. Checked 14 October 2026.
  6. Jiajun Wu | Computer Science cs.stanford.edu. Checked 14 October 2026.
  7. Congratulations to the #IJCAI2026 award winners - ΑΙhub aihub.org. Checked 14 October 2026.
  8. IJCAI Awards | IJCAI ijcai.org. Checked 14 October 2026.