Birs- 24w5308: New Directions in Machine Learning Theory

  1. Title:
    What governs predictive disparity in modern machine learning applications?

    Speaker:

    Link:

  2. Title:
    Prediction-Powered Inference

    Speaker:

    Link:

  3. Title:
    Bursting the Filter Bubble: Disincentivizing Echo Chambers in Social Networks

    Speaker:

    Link:

  4. Title:
    A Poincaré Inequality and Consistency Results for Signal Sampling on Large Graphs

    Speaker:

    Link:

  5. Title:
    Online matching with graph neural networks

    Speaker:

    Link:

  6. Title:
    Step-by-Step Diffusion

    Speaker:

    Link:

  7. Title:
    Dueling over dessert, mastering the art of repeated cake cutting

    Speaker:

    Link:

  8. Title:
    Behavioral Economics-Inspired Multi-Agent Learning

    Speaker:

    Link:

  9. Title:
    Theory of Multi-objective Machine Learning

    Speaker:

    Link:

  10. Title:
    A multigroup perspective to go beyond loss minimization in ML

    Speaker:

    Link:

  11. Title:
    Mechanisms of LLM Generalization: A Computational Approach

    Speaker:

    Link:

  12. Title:
    Efficiently learning instance-optimal linear system solvers

    Speaker:

    Link:

  13. Title:
    Should we predict in Latent Space in Self-Supervised Learning?

    Speaker:

    Link:

  14. Title:
    Towards Theoretical Understanding of Extrapolation in Data Science

    Speaker:

    Link:

  15. Title:
    Recent progress on interpretable clustering

    Speaker:

    Link:

  16. Title:
    Fair Secretaries with Unfair Predictions

    Speaker:

    Link:

  17. Title:
    Algorithmic tools for targeting sortition ideals

    Speaker:

    Link:

  18. Title:
    Majority-of-Three: The Simplest Optimal Learner?

    Speaker:

    Link:

  19. Title:
    Inherent Limitations for Characterizing Distribution Learning

    Speaker:

    Link:

  20. Title:
    VC Theory vs. Empirical DP

    Speaker:

    Link: