Snigdha Panigrahi
390 Serra Mall, Sequoia Hall,
Stanford, CA, 94305.
email : snigdha@stanford.edu

I am a statistician interested in developing methodology driven by scientific applications. My work spans across high dimensional inference, Bayesian methods, non-parametric statistics, random fields and causal inference. I explore open questions in statistical genomics that directly motivate most of my work.

I am expected to complete my PhD thesis under the supervision of Jonathan Taylor in 2018. My thesis focuses on the development of Bayesian inferential tools in the emerging area of selective inference. My research enables a practitioner to provide credible intervals and estimates like the posterior mean and MAP post selection that are based on an adjusted posterior. I am curently working with Chiara Sabatti to identify causal variants and use my methods to provide selective-bias free effect size estimates in genomics.

My other research interests involve study of multivariate random fields, characterizing path properties and establishing geometric connections. More recently, I am collaborating with Guido W.Imbens to build methods for providing and assessing credibility of causal estimates. In the past, I have collaborated with Technicolor Research, Los Altos to construct temporally evolving user personas and models for prediction of user Clickthrough Rates (CTR) based on persona features.


Publications and Preprints

Selective Inference

Snigdha Panigrahi and Jonathan Taylor. Scalable methods for Bayesian selective inference. 2017. [arxiv] [code]

Snigdha Panigrahi, Jelena Markovic and Jonathan Taylor. An MCMC-free approach to post-selective inference. 2017. [arxiv] [code]

Xiaoying Tian Harris, Snigdha Panigrahi, Jelena Markovic, Nan Bi and Jonathan Taylor. Selective sampling after solving a convex problem. 2016. [arxiv]

Snigdha Panigrahi, Jonathan Taylor and Asaf Weinstein. Bayesian Post-Selection Inference in the Linear Model. 2016. [arxiv] [code]


Random Fields

Snigdha Panigrahi, Parthanil Roy and Yimin Xiao. Maximal Moments and Uniform Modulus of Continuity for Stable Random Fields. 2017. [arxiv]

Snigdha Panigrahi, Jonathan Taylor and Sreekar Vadlamani. Kinematic Formula for Heterogeneous Gaussian Related Fields. 2017. [arxiv]


Machine Learning methods for Streaming Data

Snigdha Panigrahi and Nadia Fawaz. A relevance-scalability-interpretability tradeoff with temporally evolving user personas. 2017. [arxiv]


Statistical genomics

Snigdha Panigrahi, Jason Zhu and Chiara Sabbati. Selection-adjusted inference: an application to confidence intervals for cis-eQTL effect sizes. 2017. [arxiv]

Packages for Selective Inference

Bayesian selective inference [Github]

      This directory contains an MCMC sampler that provides Bayesian selective inference, that is based on a selection-adjusted posterior. The methods include implementation of inference (credible intervals and selection-adjusted MAP) post LASSO, forward stepwise regression, marginal screening, multi-stage screening algorithms.



MCMC free frequentist selective inference [Github]

      This directory contains code that provides selection-adjusted frequentist inference based on a a convex approximation for an intractable pivot. This also includes computation of the selective MLE that maximizes an approximate selection-adjusted law. Tools include inference post LASSO, marginal screening, forward stepwise regression and analysis of HIV data.


I have been a Teaching Assistant for the following courses at Stanford University

Stats 204: Sampling   with   Rajarshi Mukherjee (2016-17 Spring).

Stats 110: Statistical Methods in Engineering and the Physical Sciences   Timothy Patrick Daley (2016-17 Fall).

Stats 315A: Modern Applied Statistics: Learning   with   Trevor Hastie (2015-16 Winter).

Stats 110: Statistical Methods in Engineering and the Physical Sciences   with   Bala Rajaratnam (2015-16 Fall).

Stats 300C: Theory of Statistics   with   Emmanuel Candes (2014-15 Spring).

Stats 310B: Theory of Probability   with   Amir Dembo (2014-15 Winter).

Stats 300A: Theory of Statistics   with   Lester Mackey (2014-15 Fall).

Stats 60: Introduction to Statistical Methods: Precalculus   with   Guenther Walter (2013-14 Fall).

Stats 116: Theory of Probability   with   Dominique Guillot (2013-14 Summer).

Stats 60: Introduction to Statistical Methods: Precalculus   with   Jonathan Taylor (2013-14 Spring).

Invited Talks and Presentations

An approximation based approach for randomized conditional inference- with an application in eQTLs: Statistics Seminar, CMU Statistics and Data Science. 2018

An approximation based approach for randomized conditional inference- with an application in eQTLs: Statistics Seminar, University of Michigan. 2018

An approximation based approach for randomized conditional inference- with an application in eQTLs: Department of Biostatistics, Harvard T.H. Chan School of Public Health. 2018

An approximation based approach for randomized conditional inference- with an application in eQTLs: Statistics Seminar, Columbia University. 2018

Selection adjusted estimation of effect sizes with an application in eQTLs: Workshop in Operations Research and Data Science, The Fuqua School of Business, Duke University. 2017 [Conference]

Selection adjusted methods for effect size estimation in eQTLs: Statistics and genomics seminar, UC Berkeley. 2017 [Talk]

A Pseudo-likelihood approach to Selective Inference: Workshop on Higher-Order Asymptotics and Post-Selection Inference, Washington University, St. Louis. 2017 [Talk]

Bayesian Selective Inference in Linear Model: 10th International Conference on Multiple Comparison Procedures, UC Riverside. 2017 [Talk]

Bayesian Selective Inference in Linear Models: Workshop on Higher-Order Asymptotics and Post-Selection Inference, Washington University, St. Louis. 2016 [Poster]

Maximal Moments and Modulus of Continuity of Stable Random Fields: Extreme Value Analysis Conference, University of Michigan, Ann Arbor. 2015 [Talk]

Moments of Partial Maxima of Symmetric Stable Processes: Workshop on Heavy Tailed Distributions and Extreme Value Theory, Indian Statistical Institute, Kolkata. 2013 [Talk]