Schedule:

Sunday, 7 December

Structured discussion

We have invited three domain experts (health sciences, earth/climate sciences, and economics). Each keynote (35 minutes, including time for clarifying questions) will:

1. describe problems or datasets of causal interest,
2. identify obstacles in applying existing causal methods, and
3. pose a challenge or open question to stimulate discussion.

After each keynote, participants will spend 40 minutes in breakout groups engaging in group and meta-group discussions.

Finding your group

At the beginning of the workshop, you’ll receive a slip of paper. The number on the left assigns you to a meta-group, and each of the three colored shapes assigns you to a group for each keynote. The meeting point for this meta-group is indicated by a numbered sign taped to a wall inside the auditorium. Below that meta-group sign, there are three colored shape signs indicating the group.

Suppose you have the following assignment slip:

After the first keynote, you would head over to the part of the wall labeled with “4”, specifically under the red square sign there. You’d be in the same place for the second keynote; and then for the third, you’d be under the blue circle sign of the “4” meta-group.

Ask an organizer or helper if you still don’t have an assignment slip when it’s time to split into groups.

Group discussions

You’ll spend 20 minutes discussing and creating a poster with your three other group members (same meta-group number and same colored shape for the corresponding keynote). We’ll provide a large paper and some markers. Use this time to:

1. understand the problem/obstacle/challenge presented in the keynote, writing it in your own words as group, and
2. think of how to approach the problem: how do you formalize it? what are the relevant existing methods? what are concrete next steps?
3. write a summary of this approach on the poster, and be prepared to briefly present it.

Meta-group discussions

You’ll spend 15 minutes discussing in your meta-group, 5 minutes per group: The red square group will present first, then the green triangle, followed by the blue circle.

If you’d like to continue working on this problem, write your name and contact email on the back of your poster—we will collect them and follow up after the workshop.

Spotlight talks

• Causal Discovery on Galactic Star Clusters: Tools, Assumptions, and Traps
  • Mario Pasquato
• Causal Discovery with Neural Networks Applied to Centennial CO2 Fluctuations During the Last Glacial
  • Svenja Frey, Markus Jochum, Qi-Fan Wu, Roman Nuterman
• Beyond Correlational Prediction: Challenges of Causal Machine Learning in Building Energy Applications
  • Fuyang Jiang, Hussain Kazmi
• Toward Scalable and Valid Conditional Independence Testing with Spectral Features
  • Alek Fröhlich, Vladimir R Kostic, Karim Lounici, Daniel Perazzo, Massimiliano Pontil
• Robust Variance Estimation for Misspecified Causal Models via Reweighting
  • Johan de Aguas
• Confounder Mitigation with Adversarial Loss Realized via Deep Neural Networks: A Case Study on Age Predictor based on Transcriptomic Data
  • Debdas Paul, Elisa Ferrari, Alessandro Cellerino
• Functional Mendelian randomization for capturing temporal dynamics of causal effects of exposures on disease risk
  • Nicole Fontana, Francesca Ieva, Emanuele Di Angelantonio, Piercesare Secchi
• Causal Effects of Price Shocks on Food Insecurity
  • Jordi Cerda-Bautista, Vasileios Sitokonstantinou, Homer Durand, Gherardo Varando, Gustau Camps-Valls
• Invariant Learning for Robust Yield Predictions
  • Georgios Athanasiou, Nathan Mankovich, Gherardo Varando, Inti Luna, Chen Ma, Cristina Radin, Homer Durand, Jordi Cerda-Bautista, Muhammad Shoaib, Ioannis Papoutsis, Gustau Camps-Valls, Vasileios Sitokonstantinou
• Multi-View Causal Discovery without Non-Gaussianity: Identifiability and Algorithms
  • Ambroise Heurtebise, Omar Chehab, Pierre Ablin, Alexandre Gramfort, Aapo Hyvarinen
• CausalProfiler: Towards Rigorous and Transparent Evaluation in Causal Machine Learning
  • Panayiotis Panayiotou, Audrey Poinsot, Alessandro Leite, Nicolas Chesneau, Marc Schoenauer, Özgür Şimşek
• How to choose a causal effects estimator: A case example on the assessment of the gender pay gap at the company level.
  • Carlos Ameal, Jan de Neve, Tom Loeys
• Putting causal methods under a microscope: Using molecular biomarkers to generate causal knowledge and using causal methods to strengthen analyses of molecular biomarkers
  • Jennifer J. Adibi
• Counterfactual Policy Learning for Personalized Treatment Using SGLT2 Inhibitors
  • Dániel Sándor, Ákos Németh, Peter Antal
• Nonparanormal Adjusted Marginal Inference
  • Susanne Dandl, Torsten Hothorn

Poster session

• DoCluster: From Algebraic Mutations to Causal Representation Learning in Computational Biology
  • Aditya Raj Dash
• Missingness-MDPs: Bridging the Theory of Missing Data and POMDPs
  • Joshua Wendland, Markel Zubia, Roman Andriushchenko, Maris F. L. Galesloot, Milan Ceska, Henrik von Kleist, Thiago D. Simão, Maximilian Weininger, Nils Jansen
• GRITS: Iterative Graphical Imputation for Multivariate Time Series
  • Samuel Joray, Yanke Li, Diego Paez-Granados
• A Structural Extrapolation Principle for Correcting Biased CATE Estimates
  • Muneeb Aadil, Jialin Yu, Andreas Koukorinis, Nicolo Colombo, Ricardo Silva
• Nonparanormal Adjusted Marginal Inference
  • Susanne Dandl, Torsten Hothorn
• Hateful signals and where to find them
  • Sarah Masud
• Bounding Nested Counterfactuals without Structural Assumptions: An Application to Fairness Analysis
  • Eric Rossetto, Alessandro Antonucci
• CounterfactualVision: Evaluating Causal Reasoning in Vision–Language Models
  • Gokul Srinath Seetha Ram
• Physics-Informed Causal Discovery: Practical Barriers and Lessons from Real Scientific Data
  • Mahule Roy, Subhas Roy
• Replicated blood-based biomarkers for myalgic encephalomyelitis not explicable by inactivity
  • Sjoerd V. Beentjes, Artur Miralles Méharon, Julia Kaczmarczyk, Amanda Cassar, Gemma Louise Samms, Nima S. Hejazi, Ava Khamseh, Chris P. Ponting
• Counterfactual Policy Learning for Personalized Treatment Using SGLT2 Inhibitors
  • Dániel Sándor, Ákos Németh, Peter Antal
• Amortized Causal Discovery with Prior Fitted Networks
  • Mateusz Odrowaz-Sypniewski, Mateusz Olko, Mateusz Gajewski
• Policy Instruments: Graphical Methods for Instrumental Variables in Time Series
  • Henry Schöller, Søren Wengel Mogensen
• Proximal Variables Regression with Hinge Nodes: Nuisance Adjusted Bridges
  • Hubert Marek Drazkowski
• When Data Gaps Distort Causal Claims: Lessons from India’s MGNREGA Evaluation
  • Ashray Gupta, Stella Verkijk
• Putting causal methods under a microscope: Using molecular biomarkers to generate causal knowledge and using causal methods to strengthen analyses of molecular biomarkers
  • Jennifer J. Adibi
• How to choose a causal effects estimator: A case example on the assessment of the gender pay gap at the company level.
  • Carlos Ameal, Jan de Neve, Tom Loeys
• OncoSynth: Synthetic data generation for treatment effect estimation in oncology
  • Julian Welzel, Octavia-Andreea Ciora, Harry Amad, Thomas Callender, Mihaela van der Schaar, Stefan Feuerriegel
• A Generalised Kernel-based Conditional Independence Test for Causal Discovery
  • Luca Bergen, Vanessa Didelez
• Causal Counterfactuals
  • Kriti Mahajan
• CausalFairness: An Open Source Python Library for Causal Fairness Analysis
  • Kriti Mahajan
• Context-specific Causal Discovery for the Sciences - Evaluation and an Example from Climate Science
  • Oana-Iuliana Popescu, Wiebke Günther, Jakob Runge
• What Background Knowledge Can and Cannot Do For Observational Causal Discovery
  • Alexander Gilbert Reisach
• Beyond Causal Accuracy: Evaluating Representation Quality in Deep Treatment Effect Estimation
  • Ahmad Saeed Khan, Erik Schaffernicht, Johannes A. Stork
• A Divide-and-Conquer Framework for Scalable Causal Discovery with Large Language Models
  • Louis Hernandez, Alessandro Leite, Cecilia Zanni-Merk, Matthieu Boussard
• CausalProfiler: Towards Rigorous and Transparent Evaluation in Causal Machine Learning
  • Panayiotis Panayiotou, Audrey Poinsot, Alessandro Leite, Nicolas Chesneau, Marc Schoenauer, Özgür Şimşek
• Comparison of Propensity Models Based on E value
  • Cyril Biji
• Causal Regime Detection in Energy Markets With Augmented Time Series Structural Causal Models
  • Dennis Thumm
• Towards Causal Biomarkers: Characterizing and Identifying Brain Interactions for Neurological Disorders
  • Jutika Borah, Abdur R Fayjie, Rajkumar Saini, Debarun Chakarborty, Bhabesh Deka, Foteini Simistira Liwicki
• Multi-View Causal Discovery without Non-Gaussianity: Identifiability and Algorithms
  • Ambroise Heurtebise, Omar Chehab, Pierre Ablin, Alexandre Gramfort, Aapo Hyvarinen
• Heterogeneous Treatment Effect in Time-to-Event Outcomes: Harnessing Censored Data with Recursively Imputed Trees
  • Malka Gorfine, Tomer Meir, Uri Shalit
• Case Study: Learning Interventionally Consistent Surrogates for Climate Adaptation
  • Francisco Madaleno, Serio Agriesti, Carlos Lima Azevedo, Francisco C. Pereira
• Single Proxy Identifiability of Causal Effects
  • Silvan Vollmer, Niklas Pfister, Sebastian Weichwald
• Invariant Learning for Robust Yield Predictions
  • Georgios Athanasiou, Nathan Mankovich, Gherardo Varando, Inti Luna, Chen Ma, Cristina Radin, Homer Durand, Jordi Cerda-Bautista, Muhammad Shoaib, Ioannis Papoutsis, Gustau Camps-Valls, Vasileios Sitokonstantinou
• offline_rl_ope: A Python package for off-policy evaluation of offline RL models with real world data
  • Joshua William Spear, Rebecca Pope, Matthieu Komorowski, Neil J. Sebire
• Child maltreatment, adult partner violence, and young onset dementia in the UK Biobank: Sequential causal mediation analysis
  • Marie-Céline Schulte, Nicholas Magill, Karen Devries
• Causal Effects of Price Shocks on Food Insecurity
  • Jordi Cerda-Bautista, Vasileios Sitokonstantinou, Homer Durand, Gherardo Varando, Gustau Camps-Valls
• A Fast Kernel-based Conditional Independence test with Application to Causal Discovery
  • Oliver Schacht, Biwei Huang
• Open Causal: a FAIR platform for causal graphs
  • Hüseyin Küçükali
• Counterfactual Inference In Nonseparable and High-Dimensional Outcome Models Using Instrumental Variables
  • Marc Braun
• Functional Mendelian randomization for capturing temporal dynamics of causal effects of exposures on disease risk
  • Nicole Fontana, Francesca Ieva, Emanuele Di Angelantonio, Piercesare Secchi
• How Remote Contact Affected Loneliness During Lockdown: A Fully Doubly Robust Workflow with Machine Learning
  • Gaetano Tedesco
• Causal Discovery Under Missingness: Comparing Imputation Strategies and Test Statistic Stability
  • Leona Odole, Alexander Marx, Stefan Harmeling
• Debiasing Machine Learning Predictions for Causal Inference Without Additional Ground Truth Data: “One Map, Many Trials” in Satellite-Driven Poverty Analysis
  • Markus Bo Pettersson, Connor Thomas Jerzak, Adel Daoud
• Longitudinal causal inference using machine learning methods
  • Mercy Rop
• Confounder Mitigation with Adversarial Loss Realized via Deep Neural Networks: A Case Study on Age Predictor based on Transcriptomic Data
  • Debdas Paul, Elisa Ferrari, Alessandro Cellerino
• Robust Variance Estimation for Misspecified Causal Models via Reweighting
  • Johan de Aguas
• A Generalized Propensity Score Estimation Methodology for Discrete and Continuous Treatments
  • Felipe Lourenço Angelim Vieira, Alessandro Leite
• Addressing Instrument-Outcome Confounding in Instrumental Variable Estimation through Representation Learning
  • Shimeng Huang, Matthew R Robinson, Francesco Locatello
• Identifying Causal Sources from Missing Data
  • Osman Mian, Jens Kleesiek, Michael Kamp
• Towards Data-Driven Catalysis: Setting the Ground for the Application of Causality in Catalyst Design
  • Bárbara Alves Sequeira, Pedro S. F. Mendes, Adèle H. Ribeiro, Carmen Baracariza, Catarina Barata
• Soft Calibration Without Experiments: Proxy Variables for Mix Modeling
  • Felipe Lourenço Angelim Vieira
• Toward Scalable and Valid Conditional Independence Testing with Spectral Features
  • Alek Fröhlich, Vladimir R Kostic, Karim Lounici, Daniel Perazzo, Massimiliano Pontil
• Towards Testable Causal Modelling
  • Benedikt Höltgen, Robert Williamson
• Efficient Greedy Equivalence Search for Non-Score-Equivalent Criteria using Sampling
  • Rafailia Chatzianastasiou, Osman Mian, Jilles Vreeken
• A real-world case study in debiasing the application of offline reinforcement learning for policy optimisation of treatment recommendations
  • Joshua William Spear, Neil J. Sebire, Rebecca Pope, Matthieu Komorowski
• Beyond Correlational Prediction: Challenges of Causal Machine Learning in Building Energy Applications
  • Fuyang Jiang, Hussain Kazmi
• Causal Discovery with Neural Networks Applied to Centennial CO2 Fluctuations During the Last Glacial
  • Svenja Frey, Markus Jochum, Qi-Fan Wu, Roman Nuterman
• Graph learning via integer programming
  • Lucas Kook, Søren Wengel Mogensen
• Identifiable Latent Bandits: Leveraging observational data for personalized decision-making
  • Ahmet Zahid Balcıoğlu, Newton Mwai, Emil Carlsson, Fredrik D. Johansson
• Causal Discovery on Galactic Star Clusters: Tools, Assumptions, and Traps
  • Mario Pasquato