Weakly Supervised Multi-Embeddings Learning of Acoustic Models

Abstract

We trained a Siamese network with multi-task same/different information on a speech dataset, and found that it was possible to share a network for both tasks without a loss in performance. The first task was to discriminate between two same or different words, and the second was to discriminate between two same or different talkers.

Cite

Text

Synnaeve and Dupoux. "Weakly Supervised Multi-Embeddings Learning of Acoustic Models." International Conference on Learning Representations, 2015.

Markdown

[Synnaeve and Dupoux. "Weakly Supervised Multi-Embeddings Learning of Acoustic Models." International Conference on Learning Representations, 2015.](https://mlanthology.org/iclr/2015/synnaeve2015iclr-weakly/)

BibTeX

@inproceedings{synnaeve2015iclr-weakly,
  title     = {{Weakly Supervised Multi-Embeddings Learning of Acoustic Models}},
  author    = {Synnaeve, Gabriel and Dupoux, Emmanuel},
  booktitle = {International Conference on Learning Representations},
  year      = {2015},
  url       = {https://mlanthology.org/iclr/2015/synnaeve2015iclr-weakly/}
}