Unsupervised Neural Hidden Markov Models with a Continuous latent state space
Machine Learning
2021-06-14 v1
Abstract
We introduce a new procedure to neuralize unsupervised Hidden Markov Models in the continuous case. This provides higher flexibility to solve problems with underlying latent variables. This approach is evaluated on both synthetic and real data. On top of generating likely model parameters with comparable performances to off-the-shelf neural architecture (LSTMs, GRUs,..), the obtained results are easily interpretable.
Keywords
Cite
@article{arxiv.2106.06536,
title = {Unsupervised Neural Hidden Markov Models with a Continuous latent state space},
author = {Firas Jarboui and Vianney Perchet},
journal= {arXiv preprint arXiv:2106.06536},
year = {2021}
}