Samuel Kaski

Samuel Kaski

Professor of Artificial Intelligence

Aalto University

University of Manchester

Biography

I am a Machine Learning Professor operating 50-50 in the UK and Finland. In Finland I lead the Finnish Center for Artificial Intelligence FCAI, and in the UK the Manchester Centre for AI Fundamentals AI-FUN. Part of my research group is at Aalto University and part in the University of Manchester.

Current events maybe of interest:

Research interests

  • Probabilistic modelling and Bayesian inference
  • Collaborative AI for decision making and design
  • Collaboration with many fields including health, medicine, biology, user interaction, cognitive science, neuroscience; especially when formulated as virtual simulation-based laboratories
  • Timely in ML: distribution shifts, multiple tasks and sources, generative models, experimental design, simulation-based inference, humans in the loop and theory of mind, privacy-preserving learning

Current activities:

Representative Recent Publications

Full list in Google Scholar and my group’s pages

(2024). Learning Robust Statistics for Simulation-based Inference under Model Misspecification. In NeurIPS 2023.

PDF Cite

(2023). Characterizing personalized effects of family information on disease risk using graph representation learning. In MLHC 2023.

PDF Cite

(2023). Virtual laboratories: Transforming research with AI.

PDF

(2023). Online Simulator-Based Experimental Design for Cognitive Model Selection. Computational Brain & Behavior.

Cite DOI

(2023). Compositional Sculpting of Iterative Generative Processes. In NeurIPS 2023.

PDF Cite Code Poster Slides Video

(2023). Noise-Aware Statistical Inference with Differentially Private Synthetic Data. In AISTATS 2023.

PDF Cite

(2023). Differentiable user models. In UAI 2023.

PDF Cite

(2023). Zero-shot assistance in sequential decision problems. In AAAI 2023.

Cite DOI

(2023). Toward AI assistants that let designers design. AI Magazine.

Cite DOI

(2022). Human-in-the-loop assisted de novo molecular design. Journal of Cheminformatics.

Cite DOI

(2022). Likelihood-Free Inference by Ratio Estimation. Bayesian Analysis, 17.

Cite DOI

(2022). Parallel MCMC Without Embarrassing Failures. In ICML 2022.

PDF Cite

(2022). Tackling covariate shift with node-based Bayesian neural networks. In ICML 2022.

PDF Cite

Recent & Upcoming Talks

Tekoäly ja tulevaisuus
Collaborative Machine Learning for Science
Tutorial: User modeling for cooperative AI
AI Fundamentals @Manchester and why they matter