Petrus Mikkola

Contact

• Kajokuja 8 A 1, Lohja, Finland
• petrus.mikkola@gmail.com
• +358400329055
• https://petrus-mikkola.github.io

Education

• Doctor of Science (Machine Learning), Aalto University, 2019 – 2024 (expected, defense on March 28, 2024)
• Master of Science (Mathematics/Stochastics), University of Helsinki, 2016 – 2020, GPA: 4.8/5
• Master of Social Sciences (Economics), University of Helsinki, 2015 – 2018, GPA: 4.7/5
• Bachelor of Social Sciences (Economics), University of Helsinki, 2013 – 2015, GPA: 4.9/5

Research

My research centers around probabilistic machine learning, with a particular emphasis on multi-information source and human-in-the-loop settings. The main areas of focus include normalazing flows, Bayesian optimization, elicitation (knowledge elicitation, prior elicitation, preference learning), and human-AI teaming.

Related work experience

• Researcher at the University of Helsinki (9/2023 – present). I am a (postdoctoral) researcher at the University of Helsinki, working with Assoc. Prof. Arto Klami and Asst. Prof. Luigi Acerbi. Our work focuses on normalizing flows.
• Doctoral candidate at Aalto University (5/2019 – 6/2023). I was a doctoral candidate under the supervision of Prof. Samuel Kaski in Probabilistic Machine Learning group at Aalto University. My thesis, “Humans as information sources in Bayesian optimization”, was submitted to preliminary examination on August 15, 2023.
• Research Intern at the Finnish Centre of Excellence (6/2018 – 8/2018). I studied under the supervision of Prof. Konstantin Izyurov a question concerning a branch of mathematics known as discrete complex analysis.
• Intern at Statistics Finland (5/2017 – 8/2017)
• Intern at VATT, Institute for Economic Research (5/2016 – 8/2016)
• Teaching assistant at The University of Helsinki (falls 2015, 2016 and 2017)

Publications

Mikkola, P., Todorović, M., Järvi, J., Rinke, P., Kaski, S. (2020). Projective Preferential Bayesian Optimization. In Proceedings of the 37th International Conference on Machine Learning - Volume 119. ICML’20. Proceedings, Preprint, Code.

Mikkola, P., Martinelli, J., Filstroff, L., Kaski, S. (2023). Multi-Fidelity Bayesian Optimization with Unreliable Information Sources. In Proceedings of the 26th International Conference on Artificial Intelligence and Statistics - Volume 206. AISTATS’23. Proceedings, Preprint, Code.

Mikkola, P., Martin, O. A., Chandramouli, S., Hartmann, M., Abril Pla, O., Thomas, O., Pesonen, P., Corander, J., Vehtari, A., Kaski, S., Bürkner, P-C., Klami, A. (2023). Prior knowledge elicitation: The past, present, and future. Bayesian Analysis. Advance publication, Preprint.

Khoshvishkaie, A., Mikkola, P., Murena, P-A. Kaski, S. (2023). Cooperative Bayesian optimization for imperfect agents. Machine Learning and Knowledge Discovery in Databases. Research Track: European Conference, ECML PKDD 2023. Proceedings.

Workshops and preprints

Tiihonen, A., Filstroff, L., Mikkola, P., Lehto, E., Kaski, S., Todorović, M., Rinke, P. (2022). More trustworthy Bayesian optimization of materials properties by adding human into the loop. NeurIPS 2022 Workshop AI4Mat. Workshop.

Filstroff, L., Sundin, I., Mikkola, P., Tiulpin, A., Kylmäoja, J., Kaski, S. (2021). Targeted Active Learning for Bayesian Decision-Making. Preprint.

Huang, D., Filstroff, L., Mikkola, P., Zheng, R., Kaski, S. (2022). Bayesian Optimization Augmented with Actively Elicited Expert Knowledge. Preprint.

Teaching

In 2021 and 2022, I served as a teaching assistant for the course Kernel Methods in Machine Learning. During this time, I had the privilege of supervising four Bachelor’s theses for students Pyry Satama, Sasu Bruns, Marilla Malkki, and Kee Taeyoung. Additionally, I led a course from 2015 to 2017, aimed at aiding new economics students with their mathematical studies.

Community involvement

I have acted as a reviewer for IEEE TPAMI, AISTATS2022, and AISTATS2023 (awarded top-10% reviewer).