Kirankumar Shiragur
Email: shiragur [at] stanford [dot] edu;
shiragur [at] mit [dot] edu
Link: Google scholar page, DBLP page
I am currently a postdoctoral research fellow at the Broad Institute of MIT and Harvard, where I work under the guidance of Prof. Caroline Uhler. I earned my Ph.D. from Stanford University, where I had the privilege of being co-advised by Prof. Moses Charikar and Prof. Aaron Sidford. During my doctoral studies, I held a research fellow position at the Simons Institute for the Causality program in 2022. Additionally, during the summer of 2019, I was also a research intern at Adobe Research, working under the supervision of Tung Mai and Anup Rao.
Prior to Stanford, I spent an exciting year at Microsoft Research India working with Deeparnab Chakrabarty. I did my masters from the Indian Institute of Science advised by Prof. Arnab Bhattacharyya.
My research interests lie at the intersection of Algorithms, Machine Learning, Statistics and Information theory, with a recent focus on Learning Theory and Causal Inference. A major theme of my work involves building efficient algorithms for extracting information from limited data. Complementary to this theme, I also enjoy formulating new mathematical models and practical solutions for problems in the natural and social sciences, that lie beyond the range of traditional machinery.
At Stanford, I coordinated Algorithms and Friends, through which we (theory group and other Stanford researchers) offered help in solving algorithmic questions that came up in applied research.
During my Ph.D, my research was supported by the Simons-Berkeley Research Fellowship, the Stanford Data Science Scholarship and the Dantzig-Lieberman Research Fellowship.
Working papers:
Cell Type Identification through Unique Tensor Decomposition in Single-Cell RNA-Seq Data.
Orr Ashenberg, Pau Redon Munoz, Kirankumar Shiragur, Caroline Uhler
Publications:
Causal Discovery with Fewer Independence Tests.
Kirankumar Shiragur, Caroline Uhler, Jiaqi Zhang
Under review at ICML, 2024
Testing with Non-identically Distributed Samples [Paper]
Shivam Garg, Chirag Pabbaraju, Kirankumar Shiragur, Gregory Valiant
Under review at COLT, 2024
Causal Discovery under Off-Target Interventions [Paper]
Davin Choo, Kirankumar Shiragur, Caroline Uhler
Accepted at AISTATS, 2024
Membership Testing in Markov Equivalence Classes via Independence Query Oracles [Paper]
Jiaqi Zhang, Kirankumar Shiragur, Caroline Uhler
Accepted at AISTATS, 2024 (ORAL)
Meek Separators and Their Applications in Targeted Causal Discovery [Paper]
Kirankumar Shiragur, Caroline Uhler, Jiaqi Zhang
Accepted at Neurips, 2023
Structured Semidefinite Programming for Recovering Structured Preconditioners [Paper]
Arun Jambulapati, Jerry Li, Christopher Musco, Kirankumar Shiragur, Aaron Sidford, Kevin Tian
Accepted at Neurips, 2023
Adaptivity Complexity for Causal Graph Discovery [Paper]
Davin Choo, Kirankumar Shiragur
Accepted at UAI, 2023
New metrics and search algorithms for weighted causal DAGs [Paper]
Davin Choo, Kirankumar Shiragur
Accepted at ICML, 2023
Moses Charikar, Shivam Garg, Deborah M Gordon, Kirankumar Shiragur
Abstract at ITCS, 2021 and Full version at PNAS 2023
Subset verification and search algorithms for causal DAGs [Paper]
Davin Choo, Kirankumar Shiragur
Accepted at AISTATS, 2023
Verification and search algorithms for causal DAGs [Paper]
Davin Choo, Kirankumar Shiragur, Arnab Bhattacharyya
Accepted at Neurips, 2022
On the Efficient Implementation of High Accuracy Optimality of Profile Maximum Likelihood [Paper]
Moses Charikar, Zhihao Jiang, Kirankumar Shiragur, Aaron Sidford
Accepted at Neurips, 2022
The Bethe and Sinkhorn Permanents of Low Rank Matrices and Implications for Profile Maximum Likelihood [Talk, Paper]
Nima Anari, Moses Charikar, Kirankumar Shiragur, Aaron Sidford
Accepted at COLT, 2021
Assortment Planning for Two-Sided Sequential Matching Markets [Paper]
Itai Ashlagi, Anilesh Krishnaswamy, Rahul Makhijani, Daniela Saban, Kirankumar Shiragur
Accepted at OR Journal, 2021
Reward Identification in Inverse Reinforcement Learning [Paper]
Kuno Kim, Kirankumar Shiragur, Shivam Garg, Stefano Ermon
Accepted at ICML, 2021
Fractionally Log-Concave and Sector-Stable Polynomials: Counting Planar Matchings and More [Paper]
Yeganeh Alimohammadi, Nima Anari, Kirankumar Shiragur, Thuy-Duong Vuong
Accepted at STOC, 2021
On the Competitive Analysis and High Accuracy Optimality of Profile Maximum Likelihood [Paper]
Yanjun Han, Kirankumar Shiragur
Accepted at SODA, 2021
Relative-Risk and the Assessment of School Safety in the COVID-19 Pandemic: Schools May Offer Students Shelter from the Storm [Paper]
Yeganeh Alimohammadi, Kirankumar Shiragur, Ramesh Johari, David Scheinker, Kevin Schulman, and Kristan Staudenmayer
Appeared at Health Management, Policy, and Innovation, 2021
Instance Based Approximations to Profile Maximum Likelihood [Paper]
Nima Anari, Moses Charikar, Kirankumar Shiragur, Aaron Sidford
Accepted at Neurips, 2020
A General Framework for Efficient Symmetric Property Estimation [Paper]
Moses Charikar, Kirankumar Shiragur, Aaron Sidford
Accepted at Neurips, 2019
Efficient Profile Maximum Likelihood for Universal Symmetric Property Estimation [Talk1, Talk2, Paper]
Moses Charikar, Kirankumar Shiragur, Aaron Sidford
Accepted at STOC, 2019
Graph Balancing with Two Edge Types [Paper]
Deeparnab Chakrabarty, Kirankumar Shiragur
Arxiv, 2016
How friends and non-determinism affect opinion dynamics [Paper]
Arnab Bhattacharyya, Kirankumar Shiragur
Accepted at CDC, 2015
Patents:
Item transfer control systems
Kirankumar Shiragur, Tung Thanh Mai, Anup Bandigadi Rao, Ryan A. Rossi, Georgios Theocharous, Michele Saad
U.S. Patent, 2023