Featured

Welcome

I am a Principal Researcher at Microsoft Research AI, in the Information and Data Sciences group.

My current research focuses on causal analysis of large-scale social media timelines, with the vision of making causal question-answering as fast and as common as web search. With hundreds of millions of people publicly reporting on their daily experiences, we can data mine these social media streams to better understand the common and critical situations people are in, the actions they take, and their implications.  These inferences are useful for many applications including decision support tools for individuals and analytics to support policy-makers and scientists.

Causal Inference Tutorial at KDD 2018

Hi,

I greatly enjoyed presenting our tutorial on causal inference and counterfactual reasoning, with Amit Sharma, at KDD 2018 this week. There was a lot of interest — the room was standing room only — and the questions from the audience were deep and engaging.

We’ve posted our tutorial slides. The tutorial covers a conceptual introduction to counterfactual reasoning, including potential outcomes and causal graphs; basic methods for observational studies and natural experiments; and methods for validating assumptions and refuting results.

Thanks everyone,
Emre


Photo courtesy of Paul Bennet

KDD’17 Tutorial: Limits of Social Data

I had a great time presenting our tutorial on social media biases at KDD17, with Alexandra Olteanu. Thanks to the audience for deep, thoughtful questions about research design, bias mitigation, and much more.

Here are our slides (PDF). More details about social media biases and methdological pitfalls in data analysis in our survey paper.

OSSM Tutorial: Causal inference over social media

Here are the slides for my tutorial on applying causal inference to social media timelines, presented at the Workshop on Observational Studies in Social Media (OSSM) at ICWSM 2017.  The talk emphasizes counterfactual intuitions and highlights common pitfalls.  This tutorial steps through the process of analyzing social media timelines with a focus on treatment identification, covariate featurization and outcome analysis applicable to a broad class of conditioned inference algorithms.   tutorial_kiciman_ossm17

Here are the papers I referenced in the slides:

 

Maison Keynote: Learning about Personal Experiences and their Outcomes: Analyzing Social Media as an Observational Study

Here are the slides for my talk, “Learning about Personal Experiences and their Outcomes: Analyzing Social Media as an Observational Study”. maison_keynote_kiciman_2016-02-10

I presented this as a keynote at the MAISoN 2017, WSDM workshop.

The talk references the following papers:

 

Upcoming ICWSM and WebSci papers

I’ve posted the PDFs of our upcoming ICWSM and WebSci papers, Towards an Open-Domain Framework for Distilling the Outcomes of Personal Experiences from Social Media Timelines; and Information Dissemination in Heterogeneous-Intent Networks.  Also, this week at CHI, Munmun De Choudhury presented our paper, Shifts to Suicidal Ideation from Mental Health Content in Social Media, which received an honorable mention.