Extraction of Temporal Patterns in Multi-rate and Multi-modal Datasets

Antoine Liutkus 1, 2 Umut Şimşekli 3 Taylan Cemgil 3
1 MULTISPEECH - Speech Modeling for Facilitating Oral-Based Communication
Inria Nancy - Grand Est, LORIA - NLPKD - Department of Natural Language Processing & Knowledge Discovery
2 PAROLE - Analysis, perception and recognition of speech
INRIA Lorraine, LORIA - Laboratoire Lorrain de Recherche en Informatique et ses Applications
Abstract : We focus on the problem of analyzing corpora composed of irregularly sampled (multi-rate) heterogeneous temporal data. We propose a novel convolutive multi-rate factorization model for extracting multi-modal patterns from such multi-rate data. Our model builds up on previously proposed multi-view (coupled) nonnegative matrix factor-ization techniques, and extends them by accounting for heterogeneous sample rates and enabling the patterns to have a duration. We illustrate the proposed methodology on the joint study of audiovisual data for speech analysis.
Complete list of metadatas

Cited literature [14 references]  Display  Hide  Download

https://hal.inria.fr/hal-01170932
Contributor : Antoine Liutkus <>
Submitted on : Thursday, July 2, 2015 - 3:13:09 PM
Last modification on : Sunday, November 17, 2019 - 1:12:05 AM
Long-term archiving on: Tuesday, April 25, 2017 - 11:27:11 PM

File

lva_convolutive_final.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01170932, version 1

Collections

Citation

Antoine Liutkus, Umut Şimşekli, Taylan Cemgil. Extraction of Temporal Patterns in Multi-rate and Multi-modal Datasets. International Conference on Latent Variable Analysis and Signal Separation (LVA/ICA), Aug 2015, Liberec, Czech Republic. ⟨hal-01170932⟩

Share

Metrics

Record views

354

Files downloads

377