PeopleofSuperintelligence

Researcher

Carlos G. Correa

Carlos G. Correa is a researcher whose work links human planning with computational models. He modeled resource-rational task decomposition and simplified mental representations, then shared the 2022 NeurIPS Outstanding Paper Award for training meta-learning agents with natural language and programs. Later, he used program induction to study strategies and hierarchical plans, finding that people favor reusable programs. After defending his Princeton dissertation in May 2024, he began a postdoctoral role in Marcelo Mattar's NYU Psychology lab in September 2024.

Timeline

  1. By 2024

    Texas roots and Austin college

    By 2024, his path runs from a small Texas town to college in Austin.

  2. 2020–May 2022

    Task decomposition and simplified plans

    In 2020, he coauthors “Resource-rational Task Decomposition to Minimize Planning Costs,” a model of how people decompose tasks under limited planning resources. In May 2022, he coauthors a Nature paper on controlling mental representations to simplify problems and support reasoning.

  3. 2022–2023

    NeurIPS award and task model

    In 2022, he shares the NeurIPS Outstanding Paper Award for work showing how natural-language descriptions and programs can guide meta-learning agents toward human-like inductive biases. In 2023, he and his coauthors propose that people decompose tasks to reduce planning costs while maintaining task performance.

  4. 2024

    Program-based strategy induction

    In 2024, he and his coauthors use Bayesian program induction to discover strategies, trading off each strategy's simplicity against its effectiveness.

  5. By 2024

    PhD advisers at Princeton

    By 2024, his PhD work at Princeton is advised by Nathaniel Daw and Tom Griffiths.

  6. May 2024

    Princeton dissertation defense

    In May 2024, he defends his Princeton dissertation.

  7. December 2024

    Programs for hierarchical plans

    In December 2024, he and his coauthors publish a study of hierarchical plans using program generation. Their results show that people favor programs with reuse beyond minimum-description-length predictions.

  8. Since September 2024

    Postdoctoral work at NYU

    Since September 2024, he is a postdoctoral associate in NYU Psychology's Mattar Lab, advised by Marcelo Mattar.

Career

  1. Since 2024New York UniversityPostdoctoral Associate

Education

  1. 2024Princeton UniversityPhD

From the biography.

  1. The median h-index is 48 at universities and 14 at frontier AI labsIncludes New York 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.
  2. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes New York 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. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes New York University. 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. 20% of people at companies who shape AI also hold a university role, led by Google and Stanford UniversityIncludes New York University. We estimate that 539 of the 6,766 people with a known current organization (8%) hold both a university role and a company role, including 20% of people at companies and 13% of people at universities.

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 Carlos G. Correa yet. Send us a free-licensed photo.

  1. Carlos G. Correa carlos.correa.me. Checked 20 October 2026.
  2. Resource-rational Task Decomposition to Minimize Planning Costs arxiv.org. Checked 20 October 2026.
  3. People construct simplified mental representations to plan nature.com. Checked 20 October 2026.
  4. Using Natural Language and Program Abstractions to Instill Human Inductive Biases in Machines arxiv.org. Checked 20 October 2026.
  5. NeurIPS 2022 Awards neurips.cc. Checked 20 October 2026.
  6. Humans decompose tasks by trading off utility and computational cost pmc.ncbi.nlm.nih.gov. Checked 20 October 2026.
  7. Program-Based Strategy Induction for Reinforcement Learning arxiv.org. Checked 20 October 2026.
  8. Exploring the hierarchical structure of human plans via program generation arxiv.org. Checked 20 October 2026.
  9. People | Mattar Lab mattarlab.com. Checked 20 October 2026.