Timezone: »

Disentangling Causal Effects from Sets of Interventions in the Presence of Unobserved Confounders
Olivier Jeunen · Ciarán Gilligan-Lee · Rishabh Mehrotra · Mounia Lalmas


The ability to answer causal questions is crucial in many domains, as causal inference allows one to understand the impact of interventions. In many applications, only a single intervention is possible at a given time. However, in some important areas, multiple interventions are concurrently applied. Disentangling the effects of single interventions from jointly applied interventions is a challenging task---especially as simultaneously applied interventions can interact. This problem is made harder still by unobserved confounders, which influence both treatments and outcome. We address this challenge by aiming to learn the effect of a single-intervention from both observational data and sets of interventions. We prove that this is not generally possible, but provide identification proofs demonstrating that it can be achieved under non-linear continuous structural causal models with additive, multivariate Gaussian noise---even when unobserved confounders are present. Importantly, we show how to incorporate observed covariates and learn heterogeneous treatment effects. Based on the identifiability proofs, we provide an algorithm that learns the causal model parameters by pooling data from different regimes and jointly maximising the combined likelihood. The effectiveness of our method is empirically demonstrated on both synthetic and real-world data.

Author Information

Olivier Jeunen (Amazon)

Olivier Jeunen is a Postdoctoral Scientist at Amazon. In 2021, received his PhD from the University of Antwerp with a thesis titled ``Offline Approaches to Recommendation with Online Success''. His research lies at the intersection of machine learning and information retrieval -- pursuing algorithmic advances from sound theoretical foundations. He has a track record of collaborating with prominent industrial research labs, and his recent work has been recognised with the ACM RecSys ’21 Best Student Paper Award.

Ciarán Gilligan-Lee (Spotify & University College London)

I am a scientist based in London. My research focuses on causal inference and its applications. I am currently Head of the Causal Inference Research Lab at Spotify and an Honorary Associate Professor at University College London.

Rishabh Mehrotra (Spotify Research)
Mounia Lalmas (Spotify)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors

  • 2021 : Disentangling Causal Effects from Sets of Interventions in the Presence of Unobserved Confounders »
    Olivier Jeunen · Ciaran Gilligan-Lee · Rishabh Mehrotra · Mounia Lalmas
  • 2022 Spotlight: Lightning Talks 1A-3 »
    Kimia Noorbakhsh · Ronan Perry · Qi Lyu · Jiawei Jiang · Christian Toth · Olivier Jeunen · Xin Liu · Yuan Cheng · Lei Li · Manuel Rodriguez · Julius von Kügelgen · Lars Lorch · Nicolas Donati · Lukas Burkhalter · Xiao Fu · Zhongdao Wang · Songtao Feng · Ciarán Gilligan-Lee · Rishabh Mehrotra · Fangcheng Fu · Jing Yang · Bernhard Schölkopf · Ya-Li Li · Christian Knoll · Maks Ovsjanikov · Andreas Krause · Shengjin Wang · Hong Zhang · Mounia Lalmas · Bolin Ding · Bo Du · Yingbin Liang · Franz Pernkopf · Robert Peharz · Anwar Hithnawi · Julius von Kügelgen · Bo Li · Ce Zhang
  • 2020 Poster: Model Selection for Production System via Automated Online Experiments »
    Zhenwen Dai · Praveen Chandar · Ghazal Fazelnia · Benjamin Carterette · Mounia Lalmas