Mechanica Labs
EN
Mechanica Labs

Publications

2027
  1. Mining
    Geostatistical Discrete Rate Simulation of Coupled Mine, Processing, Tailings, and Backfill Systems

    Arthur Ayestas Hilgert, Alessandro Navarra

    42nd International Symposium on the Application of Computers and Operations Research in the Mineral Industry (APCOM 2027), Montréal Accepted

    Interactive figure

2026
  1. Interpretability
    Investigating what neural networks learn to fit SO(3) multiplication

    Arthur Ayestas Hilgert, Xiangzhuo Zeng, Gabriela Moisescu-Pareja, Zhaoyue Wang, Sihui Wei, Vincent Létourneau, Gavin McCracken

    Workshop on Unifying Representations in Neural Models (UniReps 2026) Under review

    Interactive figure

  2. Interpretability
    Interpreting networks learning the Sporadic Group M11 multiplication

    Zhaoyue Wang, Arthur Ayestas Hilgert, Xiangzhuo Zeng, Gabriela Moisescu-Pareja, Sihui Wei, Gavin McCracken

    Workshop on Unifying Representations in Neural Models (UniReps 2026) Under review

    Interactive figure

  3. Interpretability
    Interpreting SO(n) Multiplication: Deep Networks Generalize by Learning an Algorithm

    Arthur Ayestas Hilgert, Xiangzhuo Zeng, Sihui Wei, Gabriela Moisescu-Pareja, Vincent Létourneau, Gavin McCracken

    International Conference on Learning Representations (ICLR 2027) Under review

    Interactive figure

  4. Interpretability
    The Form of the Weights in Deep Networks Trained on Modular Addition

    Sihui Wei, Arthur Ayestas Hilgert, Xiangzhuo Zeng, Gavin McCracken

    International Conference on Learning Representations (ICLR 2027) Under review

    Interactive figure

  5. Interpretability
    Deep neural networks divide and conquer dihedral multiplication

    Sihui Wei, Gavin McCracken, Gabriela Moisescu-Pareja, Harley Wiltzer, Doina Precup, Irina Rish, Jonathan Love

    International Conference on Machine Learning (ICML 2026)

    Interactive figure

  6. Interpretability
    On the Geometry and Topology of Representations: The Manifolds of Modular Addition

    Gabriela Moisescu-Pareja, Gavin McCracken, Harley Wiltzer, Vincent Létourneau, Colin Daniels, Doina Precup, Jonathan Love

    International Conference on Learning Representations (ICLR 2026)

    Interactive figure

  7. Interpretability
    Toward a general understanding of neural representations learned by deep neural networks on group multiplications

    Arthur Ayestas Hilgert, Sihui Wei, Doina Precup, Gabriela Moisescu-Pareja, Gavin McCracken

    Topology, Algebra, and Geometry in Data Science (TAG-DS 2026) Oral

    Interactive figure

2025
  1. Interpretability
    Uncovering a Universal Abstract Algorithm for Modular Addition in Neural Networks

    Gavin McCracken, Gabriela Moisescu-Pareja, Vincent Létourneau, Doina Precup, Jonathan Love

    Neural Information Processing Systems (NeurIPS 2025)

    Interactive figure

  2. Interpretability
    Interpreting deep neural networks trained on elementary p groups reveals algorithmic structure

    Gavin McCracken, Arthur Ayestas Hilgert, Sihui Wei, Gabriela Moisescu-Pareja, Zhaoyue Wang, Jonathan Love

    Topology, Algebra, and Geometry in Data Science (TAG-DS 2025) Flash talk

    Interactive figure

  3. Interpretability
    The Geometry and Topology of Modular Addition Representations

    Gabriela Moisescu-Pareja, Gavin McCracken, Harley Wiltzer, Vincent Létourneau, Colin Daniels, Doina Precup, Jonathan Love

    Topology, Algebra, and Geometry in Data Science (TAG-DS 2025)

    Interactive figure

  4. Interpretability
    The Representations of Deep Neural Networks Trained on Dihedral Group Multiplication

    Gavin McCracken, Sihui Wei, Gabriela Moisescu-Pareja, Harley Wiltzer, Irina Rish, Jonathan Love

    NeurIPS 2025 Workshop on Symmetry and Geometry in Neural Representations

    Interactive figure