Thematic Discussion Sessions

Unless otherwise noted, lightning talks will be held in person (MIT Kolker Room, Building 26, Room 414) and over Zoom.

Upcoming Discussions

  • IAIFI NAIRR Info/Brainstorming Session
    • Tuesday, December 17, 2024, 11:00am–12:00pm
    • Led by IAIFI Computing Committee
    • As previously shared, the IAIFI Computing Committee would like to explore whether there is an opportunity to put together an IAIFI group proposal for the National AI Research Resource (NAIRR) pilot. To that end, the Computing Committee will be holding an informational and brainstorming session to discuss what resources are available and come up with possible ideas for a proposal. Please attend if you are interested in exploring additional options for computing resources beyond the IAIFI resources.

Past Discussions

Fall 2024

  • Representation/Manifold Learning: Learned Representations
    • Friday, October 18, 2024, 2:00pm–3:00pm
    • Led by Alex Gagliano and Sam Bright-Thonney (IAIFI Fellows)
    • Many of the recent successes in AI rely on the manifold hypothesis: that most high-dimensional data lie on a lower-dimensional manifold. From transfer to contrastive learning to foundation modeling, significant effort has been devoted to methods to efficiently find and map input data to this latent space. In this Thematic Discussion Session, we’ll hear from three distinguished speakers on extracting meaningful latent representations from data from the physical sciences: Aizhan Akhmetzhanova (Self-Supervised Learning for Data Compression and Inference in Cosmology), Nate Woodward, (Product Manifold Machine Learning for Physics) and David Baek, (GenEFT: Physics-Inspired Theory of Representation Learning). After three 10-minute lightning talks, we’ll have a 30 minute open discussion/Q&A session to explore the major challenges and opportunities in this field. We encourage attendees to come with questions and insights from their own work!
    • Talk Slides (for IAIFI members only)
  • IAIFI-UMass Boston Workshop: Partnership for research and training in QUantum, Artificial intelligence, Non-equilibrium physics Theory and Applications (QUANTA)
    • Friday, November 15, 2024, 2:00pm–5:00pm
    • Rahul Kulkarni, UMass Boston and others
    • In 2024, IAIFI launched a research partnership with UMass Boston. The partnership is focused on developing physics-informed reinforcement learning approaches to solve problems in quantum science and technology; in particular, developing novel approaches for entropy-regularized RL to advance current research in quantum systems and applying entropy-regularized RL tools to solve optimization problems in quantum science and technology. The event will kick off with presentations from UMass Boston researchers on this topic. We invite IAIFI researchers with research relevant to this project to sign up to give lightning talks as well (~10 minutes each). Sign up to give a lightning talk.
    • Talk Slides (for IAIFI members only)
  • IAIFI Mini-Symposium
    • Friday, September 27, 2024, 2:00pm–3:00pm
    • Jesse Thaler (MIT), Gaia Grosso (IAIFI Fellow), Jessie Micallef (IAIFI Fellow), Lina Necib (MIT), Marin Soljacic (MIT), Nikhil Mukund (MIT), Phil Harris (MIT), Shuchin Aeron (Tufts), Thomas Harvey (IAIFI Fellow), Jim Halverson (Northeastern), and Tess Smidt (MIT)
    • The Mini Symposium will include one hour of 2-minute overview talks by IAIFI Senior Investigators and Fellows, presenting their research interests broadly. Presenters will include Jesse Thaler, Gaia Grosso, Jessie Micallef, Lina Necib, Marin Soljacic, Nikhil Mukund, Phil Harris, Shuchin Aeron, Thomas Harvey, Jim Halverson, and Tess Smidt.
    • Talk Slides (for IAIFI members only)
  • State of the IAIFI
    • Friday, September 6, 2024, 2:00pm–3:00pm
    • Jesse Thaler, IAIFI Director
    • Talk Slides (for IAIFI members only)

Spring 2024

  • Generative Modeling
    • Friday, April 19, 2024, 2:00pm–3:00pm
    • Led by Carol Cuesta-Lazaro and Denis Boyda (IAIFI Fellows)
    • In this thematic discussion, we will present three distinguished speakers who will deliver 7-minute lightning talks: Ziming Liu (Generating Generative Models from Physics), Ameya Daigavane (Symphony: Improving Autoregressive Models for 3D Molecule Generation), Ge Yang (Peeking Through A Pinhole: Foundation Priors for Robot Perception). After the lightning talks, a panel discussion on generative modeling will be conducted. The discussion will focus on open questions about applying generative modeling in physics and incorporating physical principles into generative modeling techniques. Attendees are encouraged to bring questions from their respective fields to enrich the discussion.
    • Talk Slides (for IAIFI members only)
  • Uncertainty Quantification
    • Friday, February 16, 2024, 2:00pm–3:00pm
    • Led by Gaia Grosso and Jessie Micallef (IAIFI Fellows)
    • The discussion will begin with a few flash presentations (~5 min each) by current IAIFI colleagues working on this subject. Gaia and Jessie will then lead a guided discussion to collect uncertainty quantification and robustness challenges across IAIFI and discuss pathways towards potential solutions and future work. Through this discussion, the aim is to bring together IAIFI researchers to learn from each other’s experience, build connections and possibly research collaborations–you are encouraged to join the discussion no matter your research interests or background!
    • Talk Slides (for IAIFI members only)

Fall 2023

  • IAIFI Fellows Showcase
    • Friday, October 20, 2023, 2:00pm–3:00pm
    • Carolina Cuesta-Lazaro, Alexander Gagliano, Gaia Grosso, Di Luo, and Ge Yang (IAIFI Fellows)
    • This IAIFI Lightning Talk session featured several IAIFI Fellows discussing their research interests, current projects and opportunities for collaboration!
    • Talk Slides (for IAIFI members only)

Spring 2023

  • IAIFI Mini-Symposium
    • Friday, April 14, 2023, 2:00pm–3:00pm
    • Phiala Shanahan (MIT), Cora Dvorkin (Harvard), Marin Soljacic (MIT), Di Luo (MIT), Lina Necib (MIT), Tracy Slatyer (MIT), Will Detmold (MIT), Denis Boyda (MIT), Fabian Ruehle (Northeastern), Max Tegmark (MIT), Siddharth Mishra-Sharma (MIT), Carol Cuesta-Lazaro (MIT), Daniel Eisenstein (Harvard), Lisa Barsotti/Nikhil Mukund (MIT), Jessie Micallef (MIT), Mike Williams (MIT), Ge Yang (MIT), Anna Golubeva (MIT), Pulkit Agrawal (MIT), Taritree Wongjirad (Tufts), Matt Schwartz (Harvard), Phil Harris (MIT)
    • IAIFI Senior Investigators and Fellows will each briefly present their research. This will be followed by an opportunity to hold breakouts and/or network with researchers whose interests align with yours.
    • Talk Slides (for IAIFI members only)
  • Spring 2023 Lightning Talk Session II: Geometric learning/embedding strategies
    • Friday, March 17, 2023, 2:00pm–3:00pm
    • Arthur Tsang (Harvard), Emmaouil Theodosis (Harvard), Giorgi Butbaia (UNH)
    • Arthur Tsang (Grad Student, Harvard), Finding Subhalos in Galaxy-Galaxy Strong Lensing Systems; Emmanouil Theodosis (Grad Student, Harvard), Constraining neural networks to craft representations; Giorgi Butbaia (Grad Student, UNH), Numerical Calabi-Yau metrics using Spectral neural networks
    • Talk Slides (for IAIFI members only)
  • Spring 2023 Lightning Talk Session I: Robustness/fault tolerance
    • Friday, February 17, 2023, 2:00pm–3:00pm
    • Niklas Nolte (MIT), Yin Lin (MIT), Andrew Tan (MIT)
    • Niklas Nolte (Postdoc, MIT), Lipschitz Network classification; Yin Lin (Postdoc, MIT), Contour deformation for lattice QCD; Andrew Tan (PhD student, MIT), Fault-tolerant neural networks from biologically inspired error correction codes
    • Talk Slides (for IAIFI members only)

Fall 2022

  • Fall 2022 Lightning Talk Session III: Constructing Likelihoods with Neural Nets
    • Friday, December 9, 2022, 2:00pm–3:00pm
    • Georgios Stratis (Northeastern), Ziming Liu (MIT), Eric Moreno (MIT)
    • Georgios Stratis (Postdoc, Northeastern), Sample generation for the spin-fermion model using neural networks; Ziming Liu (Grad Student, MIT), Poisson flow generative models; Eric Moreno (Grad Student, MIT), Embedded Spaces to Detect Unmodelled Gravitational-Wave Signals
    • Talk Slides (for IAIFI members only)
  • Fall 2022 Lightning Talk Session II: Physics-Motivated Gradient Descent
    • Friday, November 18, 2022, 2:00pm–3:00pm
    • John Martyn (MIT), Anna Golubeva (IAIFI Fellow), Matthew Farrell (Harvard)
    • John Martyn (Grad Student, MIT), Neural Network Quantum States for Quantum Fields; Anna Golubeva (IAIFI Fellow), Fokker-Planck Equation for the Stochastic Gradient Descent; Matthew Farrell (Postdoc Fellow, Harvard), Expressivity of deep neural networks under geometric constraints (equivariance)
    • Talk Slides (for IAIFI members only)
  • Fall 2022 Lightning Talk Session I
    • Friday, September 23, 2022, 2:00pm–3:00pm
    • Andrew Saydjari (Harvard), Harold Erbin (MIT), Zev Imani (Tufts)
    • Andrew Saydjari (Grad Student, Harvard), Marginalized Data-space Gaussian Inference for Component Separation (MaDGICS): An Application to Stellar Spectra; Harold Erbin (Postdoc, MIT), Neural networks for string field theory; Zev Imani (Grad Student, Tufts), Score-Based Generative Modeling
    • Talk Slides (for IAIFI members only)

Spring 2022

  • Spring 2022 Lightning Talk Session III: Discovering Latent Structure in Artificial and Physical Systems
    • Friday, May 13, 2022, 2:00pm–3:00pm
    • Sangeon Park (MIT), Anindita Maiti (Northeastern), Gabe Margolis (MIT)
    • Sangeon Park (Grad Student, MIT), Embedding complex physics manifold into other spaces for learning the latent structure; Anindita Maiti (Grad Student, Northeastern), Non-Gaussianity and Locality in Neural Network Field Theories; Gabe Margolis (Grad Student, MIT), Rapid Locomotion via Reinforcement Learning
    • Talk Slides (for IAIFI members only)
  • S[ring 2022 Lightning Talk Session II
    • Friday, February 25, 2022, 2:00pm–3:00pm
    • Mehmet Demirtas (Northeastern), Peter Lu (MIT), Doug Finkbeiner (Harvard), Askhunna Dogra (Harvard)
    • Mehmet Demirtas (Postdoc, Northeastern), Mass Manufacturing String Compactifications; Peter Lu (Grad Student, MIT), Discovering Conservation Laws via Manifold Learning; Doug Finkbeiner (Professor, Harvard), De-blending galaxies in crowded images; Akshunna Dogra (Post-bach, Harvard), Effective theories/mathe-physical perspectives on deep learning
    • Talk Slides (for IAIFI members only)
  • Spring 2022 Lightning Talk Session I
    • Friday, February 11, 2022, 2:00pm–3:00pm
    • Julian Urban (MIT), Keiran Llewellan (MIT), Gabriel Margolis (MIT)
    • Julian Urban (Postdoc, MIT), Reconstructing QCD spectral functions with Gaussian processes; Keiran Llewellan (Grad Student, MIT), [“Building Graph Neural Networks for measurements of the Higgs boson properties; Gabriel Margolis (Grad Student, MIT), Agile Locomotion via Model-free Learning
    • Talk Slides (for IAIFI members only)

Fall 2021

  • AI Lightning Talks
    • Friday, December 17, 2021, 2:00pm–3:00pm
    • Rumen Dangovski (MIT), Taritree Wongjirad (Tufts), Akshunna Dogra (Harvard)
    • Rumen Dangovski (Grad Student, MIT), Equivariant Contrastive Learning; Taritree Wongjirad (Assistant Professor, Tufts) and Tess Smidt (Assistant Professor, MIT), Sparse Equivariant Convolutions for Neutrino Event Classification; Rikab Gambhir (Grad Student, MIT), Can you see the shape of a jet?
    • Talk Slides (for IAIFI members only)

Spring 2021

This project is supported by National Science Foundation under Cooperative Agreement PHY-2019786. Any opinions, findings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.