Panos Toulis writes in to announce this conference:
NIPS 2017 Workshop on Causal Inference and Machine Learning (WhatIF2017)
“From ‘What If?’ To ‘What Next?’ : Causal Inference and Machine Learning for Intelligent Decision Making” — December 8th 2017, Long Beach, USA.
Submission deadline for abstracts and papers: October 31, 2017
Acceptance decisions: November 7, 2017
In recent years machine learning and causal inference have both seen important advances, especially through a dramatic expansion of their theoretical and practical domains. This workshop is aimed at facilitating more interactions between researchers in machine learning, causal inference, and application domains that use both for intelligent decision making. To this effect, the 2017 ‘What If?’ To ‘What Next?’ workshop welcomes contributions from a variety of perspectives from machine learning, statistics, economics and social sciences, among others. This includes, but it is not limited to, the following topics:
– Combining experimental control and observational data
– Bandit algorithms and reinforcement learning with explicit links to causal inference and counterfactual reasoning
– Interfaces of agent-based systems and causal inference
– Handling selection bias
– Large-scale algorithms
– Applications in online systems (e.g. search, recommendation, ad placement)
– Applications in complex systems (e.g. cell biology, smart cities, computational social sciences)
– Interactive experimental control vs. counterfactual estimation from logged experiments
– Discriminative learning vs. generative modeling in counterfactual settings
We invite contributions both in the form of extended abstract and full papers. At the discretion of the organizers, some contributions will be assigned slots as short contributed talks and others will be presented as posters.
Submission length: 2 page extended abstracts or up to 8 page full paper. At least one author of each accepted paper must be available to present the paper at the workshop.
I’m pretty sure that, in these settings, there’s not much reason to be interested in the model of zero causal effects and zero systematic error, so I hope people at this conference don’t waste any time on null hypothesis significance testing except when they are talking about how to do better.