We are happy to announce the first Bridging causal inference, reinforcement learning and transfer learning (CRT2019) workshop on Thursday, September 19th 2019 at the MIT Samberg center, 6th Floor, rooms 5-6. This workshop is part of the IBM Research's AI week, registration is free but mandatory .

Causal inference is an increasingly popular research direction, focused on discovering causal relations from data and exploiting them to predict the effect of actions/interventions in a system. Recently, there has been exciting new works pointing at connections between causal inference and two other important fields of machine learning: reinforcement learning and transfer learning. Although there seems to be a natural connection between these fields, the different research communities are still separate, a situation complicated by the different terminologies and assumptions. In this workshop we will bring together these different communities in the context of causal inference. We will include speakers from academia and industry that have been pioneering research in these intersections.

Schedule

9:00-9:20 AM Breakfast and poster setup
9:20-9:30 AM Introduction
9:30-10:30 AM Elias Bareinboim (Columbia): Towards Causal Reinforcement Learning
10:30-11:00 AM Coffee and poster session
11:00-12:00 AM David Sontag (MIT): Checking Assumptions in Causal Inference from Observational Data
12:00-1:00 PM Lunch (not provided) and poster session
1:00-2:00 PM Kun Zhang (CMU): Causality, independence, and adaptive prediction
2:00-3:00 PM Coffee and poster session

Invited Speakers

Posters

  • Mark Feblowitz, Oktie Hassanzadeh and Kavitha Srinivas: Extracting Causal Knowledge from Text to Automate Planning Domain Creation
  • Xinkun Nie, Emma Brunskill and Stefan Wager: Learning When-to-Treat Policies
  • Yue Yu, Jie Chen, Tian Gao and Mo Yu: DAG-GNN: DAG Structure Learning with Graph Neural Networks
  • Debarun Bhattacharjya, Karthikeyan Shanmugam, Tian Gao, Nicholas Mattei, Kush Varshney and Dharmashankar Subramanian: Event-Driven Continuous Time Bayesian Networks
  • Kristjan Greenewald, Dmitriy Katz, Karthikeyan Shanmugam, Sara Magliacane, Murat Kocaoglu, Enric Boix Adsera and Guy Bresler: Sample Efficient Active Learning of Causal Trees
  • Tian Gao and Dennis Wei: Parallel Bayesian Network Structure Learning
  • Fredrik Johansson, Dennis Wei, Michael Oberst, Tian Gao, Gabriel Brat, David Sontag and Kush Varshney: Characterization of Overlap in Observational Studies
  • Sara Magliacane, Thijs van Ommen, Tom Claassen, Stephan Bongers, Philip Versteeg, Joris M. Mooij: Domain Adaptation by Using Causal Inference to Predict Invariant Conditional Distributions
  • Murat Kocaoglu, Karthikeyan Shanmugam, Amin Jaber and Elias Bareinboim: Characaterization and Learning of Causal Graphs with Latent Variables from Soft Interventions
  • Anish Agarwal, Romain Cosson, Devavrat Shah and Dennis Shen: Bridging Randomized Control Trials and Personalized Treatments
  • Akanksha Atrey, Kaleigh Clary and David Jensen: Evaluating Saliency Maps for Deep RL Using Counterfactual Reasoning
  • Chandler Squires, Daniel Bernstein, Caroline Uhler and Basil Saeed: Ordering-Based Learning of Causal Structures with Latent Variables
  • Zenna Tavares, James Koppel, Xin Zhang and Armando Solar-Lezama: A Language for Counterfactual Generative Models
  • Rahul Singh and Liyang Sun: De-biased Machine Learning for Compliers
  • Michael Oberst and David Sontag: Counterfactual Policy Introspection using Structural Causal Models

Organizing Committee

Sara Magliacane Karthikeyan Shanmugam Kristjan Greenewald
Murat Kocaoglu Dmitriy Katz-Rogozhnikov

Call for Posters

We invite researchers to submit work in (but not limited to) the following areas:

Submissions

Submission can be made via an EasyChair submission. Submissions of new ideas, recently published works and/or extension of existing works are welcome. Parallel submissions or submissions of under-review works are also permitted. The submission can either be a previously submitted conference or workshop paper, or an extended abstract of 1 to 3 pages (excluding references) in PDF format using NeurIPS style. Author names do not need to be anonymized. Submissions will be accepted as poster presentations. The final versions will be posted on the workshop website (and are archival but do not constitute a proceeding). 

Key Dates

Attendance

For each accepted poster, at least one author must attend the workshop and present the poster.