K
Kai Puolamäki
Researcher at University of Helsinki
Publications - 131
Citations - 2616
Kai Puolamäki is an academic researcher from University of Helsinki. The author has contributed to research in topics: Supersymmetry & Exploratory data analysis. The author has an hindex of 26, co-authored 122 publications receiving 2259 citations. Previous affiliations of Kai Puolamäki include Helsinki Institute of Physics & Helsinki Institute for Information Technology.
Papers
More filters
Posted Content
Interpreting Classifiers through Attribute Interactions in Datasets.
TL;DR: This work presents the novel ASTRID method for investigating which attribute interactions classifiers exploit when making predictions, and shows how the found attribute partitioning is related to a factorisation of the data generating distribution.
Proceedings Article
Information Retrieval by Inferring Implicit Queries from Eye Movements
TL;DR: A new search strategy, in which the information retrieval (IR) query is inferred from eye movements measured when the user is reading text during an IR task, such that relevance predictions for a large set of unseen documents are ranked significantly better than by random guessing.
Journal ArticleDOI
Heart Rate Variability for Evaluating Vigilant Attention in Partial Chronic Sleep Restriction
TL;DR: The findings suggest that HRV spectral power reflects vigilant attention in subjects exposed to partial chronic sleep restriction, and a 3-component alertness model, containing circadian and homeostatic processes coupled with sleep inertia, respectively is studied.
Book ChapterDOI
Bayesian Solutions to the Label Switching Problem
Kai Puolamäki,Samuel Kaski +1 more
TL;DR: A fully Bayesian treatment of the permutations which performs better than alternatives and can even be used to compute summaries of the posterior samples for nonparametric Bayesian methods, for which no good solutions exist so far.
Journal ArticleDOI
A randomized approximation algorithm for computing bucket orders
TL;DR: It is shown that a simple randomized algorithm has an expected constant factor approximation guarantee for fitting bucket orders to a set of pairwise preferences.