Tomi Peltola

I'm Co-founder and Machine Learning Specialist at Digisalix.

Previously, I have worked as Senior AI Scientist at Silo.AI, as Research Scientist at Curious AI, and as Post-doctoral Researcher with Professor Samuel Kaski in the Probabilistic Machine Learning research group at Helsinki Institute for Information Technology HIIT, Department of Computer Science, Aalto University, Finland.

My academic research interests include broadly Bayesian and probabilistic modelling and computation and interactive human-in-the-loop machine learning. My PhD research was about sparse Bayesian linear models and their application in genetics and epidemiology, instructed by Professor Aki Vehtari at Department of Biomedical Engineering and Computational Science, Aalto University. For the first half of 2016, I was a visiting researcher at University College London, UK, spending most of the time with the great people at Institute of Child Health and Department of Primary Care and Population Health. During my master's studies, I also worked in the FinnDiane research group at Folkhälsan Research Center.

Notes, side projects, and other miscellanea

• 2024 TIPU & PUPU (in Finnish only) -- PICO-8 fantasy console game
• 2024 Hide-and-seek (English) / Piiloleikki (Finnish) -- PICO-8 fantasy console game
• 2020 Small Expectations — a small probabilistic programming library -- code
• 2017 Visualizing symmetric matrices using D3.js -- code
• 2016 Eye diagram visualization using D3.js -- code
• 2015 Kontrafaktuaalinen eduskunta (in Finnish only) -- javascript app
• 2014 Gaussian process regression demo -- javascript app
• 2012 Git handout -- pdf

Publications

• 2020 H. Afrabandpey, T. Peltola, J. Piironen, A. Vehtari, S. Kaski. A Decision-Theoretic Approach for Model Interpretability in Bayesian Framework. Accepted to ECML 2020 (journal track, to appear in Machine Learning). [arXiv preprint link]
• 2020 C. Wittwehr, P. Blomstedt, J. P. Gosling, T. Peltola, B. Raffael, A.-N. Richarz, M. Sienkiewicz, P. Whaley, A. Worth, M. Whelan. Artificial Intelligence for Chemical Risk Assessment. Computational Toxicology, 2020, 13, 100114. [link]
• 2019 T. Peltola, M. M. Çelikok, P. Daee, S. Kaski. Machine Teaching of Active Sequential Learners. NeurIPS 2019. [link] [arXiv preprint link] [code] [website] (Older version, with a larger focus on human-computer interaction and user modelling was titled "Modelling User's Theory of AI's Mind in Interactive Intelligent Systems"; see arXiv. This was presented at Learning by Instruction Workshop at NeurIPS 2018.)
• 2019 M. M. Çelikok, T. Peltola, S. Kaski. Agent Robustness to Negative Interruptions: A Multi-agent Learning Approach. Smooth Games Optimization and Machine Learning Workshop at NeurIPS 2019. [link (pdf)]
• 2019 H. Afrabandpey, T. Peltola, S. Kaski. Human-in-the-loop Active Covariance Learning for Improving Prediction in Small Data Sets. IJCAI 2019. [link] [arXiv preprint link]
• 2019 M. M. Çelikok*, T. Peltola*, P. Daee, S. Kaski. Interactive AI with a Theory of Mind. ACM CHI 2019 Workshop: Computational Modeling in Human-Computer Interaction. [link] [arXiv preprint link, slightly updated version]
• 2018 T. Peltola, J. Jokinen, S. Kaski. Probabilistic Formulation of the Take The Best Heuristic. Annual Meeting of the Cognitive Science Society, CogSci 2018 Proceedings. [link] [arXiv preprint link] [code]
• 2018 T. Peltola. Local Interpretable Model-Agnostic Explanations of Bayesian Predictive Models via Kullback-Leibler Projections. Extended abstract/short paper, Proceedings of the 2nd Workshop on Explainable Artificial Intelligence (XAI 2018) at IJCAI/ECAI 2018. [link]
• 2018 I. Sundin*, T. Peltola*, L. Micallef, H. Afrabandpey, M. Soare, M. M. Majumder, P. Daee, C. He, B. Serim, A. Havulinna, C. Heckman, G. Jacucci, P. Marttinen, S. Kaski. Improving genomics-based predictions for precision medicine through active elicitation of expert knowledge. Bioinformatics, 2018, 34(13):i395-i403. [link] [code]
• 2018 P. Daee*, T. Peltola*, A. Vehtari, S. Kaski. User Modelling for Avoiding Overfitting in Interactive Knowledge Elicitation for Prediction. Proceedings of the 23rd International Conference on Intelligent User Interfaces, IUI 2018, 305-310. [link] [code]
• 2018 I. Petersen, T. Peltola, S. Kaski, K. R. Walters, S. Hardoon. Depression, depressive symptoms and treatments in women who have recently given birth: UK cohort study. BMJ Open, 2018;8:e022152. [link]
• 2018 E. Marques, T. Peltola, S. Kaski, J. Klefström. Phenotype-driven identification of epithelial signalling clusters. Scientific reports, 2018, 8, 4034. [link]
• 2017 D. Smirnov, F. Lachat, T. Peltola, J.M. Lahnakoski, O.-P. Koistinen, E. Glerean, A. Vehtari, R. Hari, M. Sams, L. Nummenmaa. Brain-to-brain hyperclassification reveals action-specific motor mapping of observed actions in humans. PLoS ONE, 2017, 12(12): e0189508. [link]
• 2017 P. Daee*, T. Peltola*, M. Soare*, S. Kaski. Knowledge elicitation via sequential probabilistic inference for high-dimensional prediction. Machine Learning, 2017, 106 (9-10), 1599-1620; ECML PKDD 2017 Special Issue. [link] [code]
• 2017 H. Afrabandpey, T. Peltola, S. Kaski. Interactive prior elicitation of feature similarities for small sample size prediction. Proceedings of the 25th Conference on User Modeling, Adaptation and Personalization, UMAP 2017, 265-269. [link]
• 2017 L. Micallef*, I. Sundin*, P. Marttinen*, M. Ammad-ud-din, T. Peltola, M. Soare, G. Jacucci, S. Kaski. Interactive elicitation of knowledge on feature relevance improves predictions in small data sets. Proceedings of the 22nd International Conference on Intelligent User Interfaces, IUI 2017, 547-552. [link]
• 2014 T. Peltola. Sparse Bayesian Linear Models: Computational Advances and Applications in Epidemiology. Aalto University publication series DOCTORAL DISSERTATIONS, 206/2014. [link]
• 2014 T. Peltola, A.S. Havulinna, V. Salomaa, A. Vehtari. Hierarchical Bayesian Survival Analysis and Projective Covariate Selection in Cardiovascular Event Risk Prediction. Proceedings of the Eleventh UAI Bayesian Modeling Applications Workshop, 2014: 77–86. [link (pdf)] [code]
• 2014 T. Peltola, P. Jylänki, A. Vehtari. Expectation Propagation for Likelihoods Depending on an Inner Product of Two Multivariate Random Variables. Proceedings of the Seventeenth International Conference on Artificial Intelligence and Statistics, AISTATS 2014, PMLR 33: 769–777. [link] [code]
• 2012 T. Peltola, P. Marttinen, A. Vehtari. Finite Adaptation and Multistep Moves in the Metropolis-Hastings Algorithm for Variable Selection in Genome-Wide Association Analysis. PLoS ONE, 2012, 7(11): e49445. [link] [code]
• 2012 T. Peltola, P. Marttinen, A. Jula, V. Salomaa, M. Perola, A. Vehtari. Bayesian Variable Selection in Searching for Additive and Dominant Effects in Genome-wide Data. PLoS ONE, 2012, 7(1): e29115. [link] [code]
• 2012 S.M. Kuusisto, T. Peltola, M. Laitinen, L.S. Kumpula, V.–P. Mäkinen, T. Salonurmi, P. Hedberg, M.J. Savolainen, M.L. Hannuksela, M. Ala-Korpela. The interplay between lipoprotein phenotypes, adiponectin and alcohol consumption. Annals of Medicine, 2012, 44(5):513–22. [link]
• 2012 V.–P. Mäkinen, T. Tynkkynen, P. Soininen, C. Forsblom, T. Peltola, A.J. Kangas, P.–H. Groop and M. Ala-Korpela. Sphingomyelin is associated with kidney disease in type 1 diabetes (The FinnDiane Study). Metabolomics, 2012, 8(3): 1782–90. [link]
• 2012 V.–P. Mäkinen, T. Tynkkynen, P. Soininen, T. Peltola, A.J. Kangas, C. Forsblom, L.M. Thorn, K. Kaski, R. Laatikainen, M. Ala-Korpela and P.–H. Groop. Metabolic diversity of progressive kidney disease in 325 patients with type 1 diabetes (the FinnDiane Study). Journal of Proteome Research, 2012, 11(3): 1782–90. [link]
• 2012 D. Gordin, J. Wadén, C. Forsblom, L.M. Thorn, M. Rosengård-Bärlund, O. Heikkilä, M. Saraheimo, N. Tolonen, K. Hietala, A. Soro-Paavonen, L. Salovaara, V.–P. Mäkinen, T. Peltola, L. Bernardi, P.–H. Groop. Arterial stiffness and vascular complications in patients with type 1 diabetes: The Finnish Diabetic Nephropathy (FinnDiane) Study. Annals of Medicine, 2012, 44(2): 196–204. [link]

Teaching, communal, and related activities

• 2021 Contributed an improved gradient computation for grid_sample-function in PyTorch. See the issue on GitHub for details.
• 2021 Volunteered as TA for Bayesian Data Analysis (Global South) course organized by Professor Aki Vehtari (Aalto University).
• 2020 Gave an invited talk on Interactive AI and Machine Teaching of Active Sequential Learners in the Machine Learning Coffee Seminar series organized by the Finnish Center for Artificial Intelligence. [slides]
• 2018 Gave an invited talk on User and Machine Theory of Mind in the Understandability session of AI Day 2018 organized by the Finnish Center for Artificial Intelligence. [slides (pdf)]
• 2018 Gave lecture on Bayesian (human-in-the-loop) optimization for Computational User Interface Design course at Aalto University. [slides]
• 2018 One of the top reviewers for the thirty-second conference on Neural Information Processing Systems (NeurIPS 2018).
• 2018 Organized Probabilistic Modelling for Cognition and Interaction course at Aalto University.
• 2017-2018 Gave lecture on Survival Analysis and instructed project works on the topic for Statistical Genetics and Personalised Medicine course at Aalto University.
• 2017 Organized a reading group for the book Reinforcement Learning: An Introduction by Sutton and Barto for the Probabilistic Machine Learning research group at Aalto University.
• 2016 Co-organized a seminar course on Machine Learning and Sequential Decision Making at Aalto University.
• 2015 Organized a reading group for the book Deep Learning by Goodfellow, Bengio, and Courville for the Probabilistic Machine Learning research group at Aalto University.
• 2015 Co-organized a seminar course on Machine Learning and Differential Privacy at Aalto University.
• 2009-2013 TAed for Computational Science course at Aalto University.
• 2010 TAed for Computational Systems Biology course at Aalto University.