Publications
2025
-
- STOCBreaking the Barrier of Self-Concordant Barriers: Faster Interior Point Methods for M-Matrices57th ACM Symposium on Theory of Computing, 2025
2023
- ICMLQuantized Distributed Training of Large Models with Convergence Guarantees40th International Conference on Machine Learning, 2023
- ICLRCrAM: A Compression-Aware MinimizerInternational Conference on Learning Representations, 2023
2021
- NeurIPSAC/DC: Alternating Compressed/Decompressed Training of Deep Neural NetworksAdvances in Neural Information Processing Systems, 2021
2020
2019
2018
-
- ICLRTowards Deep Learning Models Resistant to Adversarial AttacksInternational Conference on Learning Representations, 2018Oral presentation at the Principled Approaches to Deep Learning workshop, ICML 2017
MNIST Challenge, CIFAR10 Challenge
2017
- ECMultidimensional Binary Search for Contextual Decision-MakingACM Conference on Economics and Computation, 2017Journal version in Operations Research
- STOCAlmost-Linear-Time Algorithms for Markov Chains and New Spectral Primitives for Directed Graphs49th Annual ACM SIGACT Symposium on Theory of Computing, 2017Invited to Highlights of Algorithms 2018
- SODANegative-Weight Shortest Paths and Unit Capacity Minimum Cost Flow in Õ(m10/7 log W) Time TimeTwenty-Eighth Annual ACM-SIAM Symposium on Discrete Algorithms, 2017Invited to Highlights of Algorithms 2017
2016
- mBioPhenotypic Profiling Reveals That Candida Albicans Opaque Cells Represent a Metabolically Specialized Cell State Compared to Default White CellsmBio, 2016
2015
- ICMLTight Bounds for Approximate Caratheodory and Beyond34th International Conference on Machine Learning, 2015Oral presentation at the Informs Optimization Society Conference 2016