PeopleofSuperintelligence

Researcher

Pablo Barceló

Also known as Pablo Barcelo

Pablo Barceló is a computer scientist who applies logic and database theory to machine learning, with results on graph neural networks and Transformers. After a PhD in computer science at the University of Toronto in 2006, he edited the database theory column of SIGMOD Record from 2011 to 2014 and chaired ICDT 2019 and ACM PODS 2022. His 2020 ICLR paper 'The Logical Expressiveness of Graph Neural Networks' and his 2021 JMLR paper 'Attention is Turing-Complete', coauthored with Jorge Pérez and Javier Marinkovic, examined what graph neural networks and Transformers can express and compute. In 2026, he shared the Alonzo Church Award with six other researchers for logical languages for querying property graphs that were adopted in the SQL and GQL standards. He is a full professor at Pontificia Universidad Católica de Chile and directs its Institute for Mathematical and Computational Engineering.

Timeline

  1. 2006

    PhD at Toronto

    He earns a PhD in Computer Science at the University of Toronto in 2006.

  2. 2011–2014

    SIGMOD Record theory column

    From 2011 to 2014, he edits the database theory column of SIGMOD Record.

  3. 2019–2022

    Conference chair and IMC director

    He chairs ICDT 2019 and ACM PODS 2022. In April 2022, he is deputy director of the Millennium Institute for Foundational Research on Data (IMFD), a researcher at the National Center for Artificial Intelligence (Cenia) and director of the Institute for Mathematical and Computational Engineering at Pontificia Universidad Católica (IMC UC).

  4. 2020

    Logic of graph neural networks

    In 2020, he coauthors 'The Logical Expressiveness of Graph Neural Networks' at ICLR with Egor V. Kostylev, Mikael Monet, Jorge Pérez, Juan Reutter and Juan Pablo Silva. The paper shows that adding readout functions to a popular class of GNNs called AC-GNNs lets them capture all classifiers expressible in the logic FOC2, a fragment of first-order logic.

  5. 2021

    Attention is Turing-complete

    In 2021, he coauthors 'Attention is Turing-Complete' in JMLR with Jorge Pérez and Javier Marinkovic. The paper shows that the Transformer with hard attention is Turing complete based on its capacity to compute and access internal dense representations of the data.

  6. 2024

    Transformer encoders and logic

    In 2024, he coauthors 'Logical Languages Accepted by Transformer Encoders with Hard Attention' at ICLR with Alexander Kozachinskiy, Anthony Widjaja Lin and Vladimir Podolskii. The paper shows that unique hard attention Transformer encoders can recognize all languages definable in first-order logic with arbitrary unary numerical predicates, and that average hard attention encoders handle the same logic enriched with counting terms.

  7. 2026

    An Alonzo Church Award

    In 2026, he shares the Alonzo Church Award for Outstanding Contributions to Logic and Computation with Leonid Libkin, Wim Martens, Juan Reutter, Miguel Romero, Moshe Vardi and Domagoj Vrgoč. The award recognizes foundations of logical languages for querying property graphs that were adopted in the SQL and GQL international standards, and the cited work includes his 2014 Journal of the ACM paper 'Querying Regular Graph Patterns' with Libkin and Reutter. The award is presented at ICALP 2026.

  8. As of October 2026

    Professor and IMC director

    He is a full professor at Pontificia Universidad Católica de Chile, where he also directs the Institute for Mathematical and Computational Engineering. His interests are database theory, logic in computer science and their relationship with machine learning. His research at CENIA covers the expressive power of graph neural networks, the interpretability of machine learning models and the computational power of Transformers.

Career

  1. Since 2022National Center for Artificial IntelligenceResearcher
  2. Since 2022Pontifical Catholic University of ChileDirector
  3. From 2022Millennium Institute for Foundational Research on DataDeputy Director

Education

  1. 2006University of TorontoPhD in Computer Science

From the biography.

  1. The United States employs an estimated 48% of people shaping AIIncludes National Center for Artificial Intelligence. We estimate where the people shaping AI work today, how universities, companies and research institutes split across countries, and how country shares have changed since the 1960s.
  2. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Pontifical Catholic University of Chile. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  3. The median h-index is 48 at universities and 14 at frontier AI labsIncludes University of Toronto. 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.
  4. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes University of Toronto. 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.

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 Pablo Barceló yet. Send us a free-licensed photo.

  1. Perfiles - CENIA cenia.cl. Checked 22 October 2026.
  2. Pablo Barcelo - Simons Institute for the Theory of Computing simons.berkeley.edu. Checked 22 October 2026.
  3. Pablo Barceló – Instituto Milenio Fundamentos de los Datos imfd.cl. Checked 22 October 2026.
  4. CAIML Seminar: Graph Neural Networks with Local Graph Parameters caiml.dbai.tuwien.ac.at. Checked 22 October 2026.
  5. The Logical Expressiveness of Graph Neural Networks | ML Anthology mlanthology.org. Checked 22 October 2026.
  6. Attention is Turing-Complete jmlr.org. Checked 22 October 2026.
  7. Logical Languages Accepted by Transformer Encoders with Hard Attention | ML Anthology mlanthology.org. Checked 22 October 2026.
  8. Winner of the 2026 Alonzo Church Award – ACM Special Interest Group on Logic and Computation siglog.org. Checked 22 October 2026.
  9. Investigadores chilenos ganan el Premio Alonzo Church 2026, mayor reconocimiento mundial en computación emol.com. Checked 22 October 2026.