Search Results for author: Hermanni Hälvä

Found 5 papers, 2 papers with code

Identifiable Feature Learning for Spatial Data with Nonlinear ICA

no code implementations28 Nov 2023 Hermanni Hälvä, Jonathan So, Richard E. Turner, Aapo Hyvärinen

In this paper, we introduce a new nonlinear ICA framework that employs $t$-process (TP) latent components which apply naturally to data with higher-dimensional dependency structures, such as spatial and spatio-temporal data.

Disentanglement Variational Inference

Hidden Markov Nonlinear ICA: Unsupervised Learning from Nonstationary Time Series

1 code implementation22 Jun 2020 Hermanni Hälvä, Aapo Hyvärinen

The central idea in such works is that the latent components are assumed to be independent conditional on some observed auxiliary variables, such as the time-segment index.

Disentanglement Time Series +1

Independent Innovation Analysis for Nonlinear Vector Autoregressive Process

no code implementations19 Jun 2020 Hiroshi Morioka, Hermanni Hälvä, Aapo Hyvärinen

Additivity greatly limits the generality of the model, hindering analysis of general NVAR processes which have nonlinear interactions between the innovations.

Time Series Time Series Analysis

Computational Graph Approach for Detection of Composite Human Activities

no code implementations5 Dec 2018 Niko Reunanen, Ville Könönen, Hermanni Hälvä, Jani Mäntyjärvi, Arttu Lämsä, Jussi Liikka

The resulting model learns 1) a representation of the atomic activities of a sport and 2) to classify physical activities as compositions of the atomic activities.

Action Detection Activity Detection

Cannot find the paper you are looking for? You can Submit a new open access paper.