Nicholas Haber

Timeline

1. 2008–2013

   Mathematics at Brown and Stanford

   He earns a Sc.B. in mathematics and economics at Brown University in 2008. In 2013, he completes a mathematics PhD at Stanford University under András Vasy, studying partial differential equations.

2. 2013–2014

   Postdoctoral posts and Sension

   In 2013, he becomes chief scientific officer of Sension, a company that applies computer vision to online education. He holds postdoctoral fellowships at the Mathematical Sciences Research Institute in 2013 and McGill University in 2014.

3. August 2014

   Autism research in Wall Lab

   He joins Stanford’s Wall Lab as a postdoctoral scholar in August 2014, working on social cue recognition and gaze tracking for autism therapy.

4. 2017–2018

   NeuroAILab and learning through play

   He joins Stanford’s NeuroAILab in January 2017. In 2018, he coauthors Learning to Play With Intrinsically-Motivated, Self-Aware Agents, whose neural networks learn about objects by predicting the effects of their actions and seeking interactions that challenge those predictions.

5. 2019

   The Superpower Glass trial

   A co-founder of Stanford’s Autism Glass Project, he coauthors a randomized trial of Superpower Glass in 2019. The system uses Google Glass and a smartphone app to help children with autism recognize facial expressions. The trial includes 71 children and finds an improvement on the Vineland socialization scale when the tool is added to standard therapy.

6. January 2020

   Starting the Autonomous Agents Lab

   He joins Stanford’s Graduate School of Education as an assistant professor in January 2020, as part of its Learning Differences and the Future of Special Education initiative. His Autonomous Agents Lab also starts at Stanford in January 2020.

7. 2023

   Curious Replay for adaptation

   In 2023, he coauthors Curious Replay for Model-based Adaptation with Isaac Kauvar, Chris Doyle, and Linqi Zhou. The method uses a curiosity signal to prioritize past experiences when training an agent’s world model, improving exploration and performance on the Crafter benchmark.

8. 2024

   Minimo learns formal mathematics

   With Gabriel Poesia, David Broman, and Noah D. Goodman, he coauthors Learning Formal Mathematics From Intrinsic Motivation in 2024. Their Minimo agent learns to propose mathematical conjectures and prove them, starting from axioms in logic, arithmetic, and group theory.

9. 2025

   Testing chatbots for therapy

   In 2025, he coauthors a study that compares large language models with the best practices of human mental health professionals. The team tests chatbots with portions of real therapy transcripts; in one test, a chatbot supplies a list of bridges over a given height after a user describes losing a job.

10. 2026

    Research and teaching at Stanford

    In 2026, he teaches Interactive and Embodied Learning at Stanford, covering memory, play, and social learning. He is an assistant professor at the Stanford Graduate School of Education and, by courtesy, in Computer Science. His group develops AI systems that model early learning through play, social interaction, and curiosity.