publications

2025

  1. NeurIPS’25
    Understanding the Gain from Data Filtering in Multimodal Contrastive Learning
    Divyansh Pareek, Sewoong Oh, and Simon S Du
    Advances in Neural Information Processing Systems, 2025
    PDF

2024

  1. NeurIPS’24
    Understanding the Gains from Repeated Self-Distillation
    Divyansh Pareek, Simon S Du, and Sewoong Oh
    Advances in Neural Information Processing Systems, 2024
    arXiv

2022

  1. ICLR’22
    The Effects of Invertibility on the Representational Complexity of Encoders in Variational Autoencoders
    Divyansh Pareek, and Andrej Risteski
    International Conference on Learning Representations, 2022
    arXiv

2018

  1. ESA’18
    Strong Collapse for Persistence
    Jean-Daniel Boissonnat, Siddharth Pritam, and Divyansh Pareek
    European Symposium on Algorithms, 2018
    arXiv