Fredrik D. Johansson
Fredrik D. Johansson is a machine-learning researcher whose work connects graph learning, causal inference and healthcare decision-making. He completed a licentiate thesis on graph kernels at Chalmers in 2014 and a PhD there, with a 2016 doctoral thesis on geometric graph embeddings; he also coauthored work on counterfactual inference and treatment effects. After a postdoctoral appointment at MIT, he joined Chalmers in 2019, where the Healthy AI Lab began in 2020 under his leadership. He presented a causal-effects talk as an assistant professor at Chalmers in February 2022, coauthored a generalization-bounds study in 2022 and a missing-values paper in 2025, and is an associate professor at Chalmers and PI of the Healthy AI Lab as of October 2026.
Timeline
-
2013–2014
Graph kernels and graph classification
In 2013, he coauthors Entity disambiguation in anonymized graphs using graph kernels, which classifies anonymized graph nodes by graph-kernel similarity and SVMs. His 2014 licentiate thesis, Global geometric graph kernels and applications, explores graph kernels and graph classification.
-
2016
A PhD on graph embeddings
In 2016, he publishes Learning with Geometric Embeddings of Graphs, a doctoral thesis that develops graph embeddings for supervised and unsupervised tasks such as classification and clustering. He completes his PhD at Chalmers under Professor Devdatt Dubhashi, whose advising focuses on learning from graph-structured data.
-
2016–2017
Counterfactual inference and treatment effects
In 2016, he coauthors Learning Representations for Counterfactual Inference with Uri Shalit and David Sontag. The paper brings domain adaptation and representation learning together for counterfactual inference. In 2017, he coauthors an individual-treatment-effect method with Shalit and Sontag that learns balanced representations and provides generalization bounds.
-
2019–2020
A lab for healthy AI
After a postdoc with Professor David Sontag at MIT's Institute for Medical Engineering and Science, he joins Chalmers in 2019. The Healthy AI Lab starts there in 2020, with him as its principal investigator.
-
2022
Bounds for causal-effect estimates
In February 2022, he presents Generalization Bounds for Estimation of Causal Effects while an assistant professor at Chalmers. In 2022, he coauthors a JMLR paper on estimating potential outcomes and causal effects, which gives generalization bounds for estimates from observational data.
-
2025
Models that avoid missing values
In 2025, he coauthors Prediction Models That Learn to Avoid Missing Values, which introduces a missingness-avoiding framework for decision trees, tree ensembles and sparse linear models. The methods reduce reliance on features missing at test time while maintaining predictive performance.
-
2026
An ELLIS Scholar in Sweden
In 2026, ELLIS establishes its Sweden unit, where he is a Scholar at Chalmers.
-
As of October 2026
Associate professor and lab PI
As of October 2026, he is an associate professor in Data Science and AI at Chalmers and PI of the Healthy AI Lab. His research focuses on causal inference in machine learning and decision-making in healthcare.
Colleagues
Sources and credits
Photo: We have no published photo of Fredrik D. Johansson yet. Send us a free-licensed photo.
- Entity disambiguation in anonymized graphs using graph kernels
- Global geometric graph kernels and applications
- Learning with Geometric Embeddings of Graphs
- Fredrik Johansson | Healthy AI Lab
- Learning Representations for Counterfactual Inference
- Estimating individual treatment effect: generalization bounds and algorithms
- Healthy AI Lab | Healthy AI lab: Research in machine learning and healthcare
- Causal Inference and Missing Data Group at Inria
- Generalization Bounds and Representation Learning for Estimation of Potential Outcomes and Causal Effects
- [2505.03393] Prediction Models That Learn to Avoid Missing Values
- Fredrik D. Johansson | European Laboratory for Learning and Intelligent Systems
- Sweden | European Laboratory for Learning and Intelligent Systems
- Chalmers Research: Fredrik Johansson
- Fredrik D. Johansson, Associate Professor in Machine Learning at Chalmers University