About me

I am a final-year PhD candidate at TU Delft supervised by Jesse Krijthe and Marcel Reinders in the Pattern Recognition Laboratory. As part of my PhD, I was also a visiting graduate student at Harvard University, working with Issa Dahabreh in the CAUSALab. Morever, I spent some time as a machine learning scientist intern at Booking.com in Amsterdam.

I am originally from Sweden, where I earned my BSc in Engineering Physics and MSc in Engineering Mathematics from Chalmers University of Technology. During my studies, I also interned at NASA Goddard Space Flight Center. My full CV can be found here.

My research interests revolve around causal inference and machine learning, and how we can better answer causal questions using data. In particular, I am interested in how we can combine datasets from different sources, such as randomized controlled trials and observational studies, using flexible machine learning models to improve the sample efficiency and validity of causal inference. Along these lines, I have most recently focused on the following two important questions:

• Trial augmentation How can we to safely improve the efficiency in randomized controlled trials using data from external sources?
• Causal falsification How can we falsify the validity of assumptions, such as the absence of unmeasured confounding, required to identify and estimate treatment effects?

If you're interested in either of these questions, please see my publications but also check out my Python package causal-falsify.

• [Jan 2026] Our paper Robust estimation of heterogeneous treatment effects in randomized trials leveraging external data accepted to AISTATS 2026.
• [Aug 2025] Our paper Robust integration of external control data in randomized trials has been accepted to Biometrics and further distinguished as a Discussion paper, an honor extended to roughly 3% of accepted articles.
• [May 2025] Giving an invited talk on trial augmentation during the Amsterdam Causal Inference Meeting.
• [May 2025] Our paper Falsification of Unconfoundedness by Testing Independence of Causal Mechanisms accepted to ICML 2025.
• [Apr 2025] Giving an oral presentation on trial augmentation during the European Causal Inference Meeting 2025 in Ghent.
• [Feb 2025] New preprint Qini curve estimation under clustered network interference.
• [Jan 2025] Became recipient of the G-Research grant for early-career researchers.
• [Oct 2023] Our paper Putting Causal Identification to the Test: Falsification using Multi-Environment Data accepted to NeurIPS 2023 Causal Representation Learning workshop.
• [Sep 2023] Our paper Detecting hidden confounding in observational data using multiple environments accepted to NeurIPS 2023.
• [Aug 2023] Our paper on benchmarking surrogate-based optimation algorithms from my masters accepted to Applied Soft Computing. You can read it here.
• [Sep 2022] Presented a poster at the ELLIS Doctoral Symposium in Alicante, Spain.
• [Jun 2022] Presented a poster at the Machine Learning Summer School in Krakow, Poland.
• [Jan 2022] Our paper Using Time-Series Privileged Information for Provably Efficient Learning of Prediction Models accepted to AISTATS 2022.
• [Jul 2021] Got 1st place on the GECCO 2021 Industrial Challenge (limited evaluation track) together with Laurens Bliek and Arthur Guijt, our approach is described here.
• [Jun 2021] Finished my thesis and graduated with a MSc in Engineering Mathematics from Chalmers University of Technology.
• [Nov 2020] Our paper, Continuous Surrogate-Based Optimization Algorithms Are Well-Suited for Expensive Discrete Problems, was accepted at BNAIC/BeneLearn 2020 and also selected for a special volume featuring the best papers from the conference (top ~30%).

Research

The full list of papers can also be found on my Google scholar profile.

Publications

• Rickard Karlsson, Piersilvio De Bartolomeis, Issa J. Dahabreh, Jesse H. Krijthe Robust estimation of heterogeneous treatment effects in randomized trials leveraging external data. AISTATS, 2026 (Forthcoming).
  [paper] [code]
• Rickard Karlsson*, Guanbo Wang*, Piersilvio De Bartolomeis, Jesse H. Krijthe, Issa J. Dahabreh Robust integration of external control data in randomized trials. Biometrics, 2026 (Forthcoming). *Equal contribution
  [paper] [code]
• Rickard Karlsson, Jesse H. Krijthe Falsification of Unconfoundedness by Testing Independence of Causal Mechanisms. ICML, 2025.
  [paper] [code]
• Rickard Karlsson, Jesse H. Krijthe Detecting Hidden Confounding in Observational Data using Multiple Environments. NeurIPS, 2023.
  [paper] [code]
• Rickard Karlsson, Ștefan Creastă, Jesse H. Krijthe Putting Causal Identification to the Test: Falsification using Multi-Environment Data. CLR Workshop, NeurIPS, 2023.
  [paper]
• Laurens Bliek, Arthur Guijt, Rickard Karlsson, Sicco Verwer, Mathijs de Weerdt Benchmarking Surrogate-based Optimisation Algorithms on Expensive Black-box Functions. Applied Soft Computing, 2023.
  [paper] [code]
• Rickard Karlsson*, Martin Willbo*, Zeshan Hussain, Rahul G. Krishnan, David Sontag, Fredrik D. Johansson Using Time-Series Privileged Information for Provably Efficient Learning of Prediction Models. AISTATS, 2022. *Equal contribution
  [paper] [code]
• Rickard Karlsson, Laurens Bliek, Sicco Verwer, Mathijs de Weerdt Continuous Surrogate-based Optimization Algorithms are Well-suited for Expensive Discrete Problems. BNAIC/Benelearn, 2020.
  [paper]

Preprints

• Rickard Karlsson*, Bram van den Akker*, Felipe Moraes, Hugo M. Proença, Jesse H. Krijthe Qini curve estimation under clustered network interference. Under review, 2025. *Equal contribution
  [paper] [code]

Software

• causal-falsify: a Python package implementing various falsification algorithms for the assumption of no unmeasured confounding when having observational data from multiple sources.

Contact

You can contact me through email at r.k.a.karlsson{at}tudelft.nl, feel free to reach out for collaborations. I am also active on Bluesky.

Office

Room 6.E.040, Building 28 at Delft University of Technology
Van Mourik Broekmanweg 6, 2628 XE Delft, The Netherlands