Micha Elsner — accepted AI-conference papers

2023

Exploring How Generative Adversarial Networks Learn Phonological Representations

ACL 2023long

This paper explores how Generative Adversarial Networks (GANs) learn representations of phonological phenomena. We analyze how GANs encode contrastive and non-contrastive nasality in French and English vowels by applying the ciwGAN architecture (Begus, 2021). Begus claims that ciwGAN encodes linguis…

Cited by 5 Source ↗PDF ↗Scholar ↗

2015

Unsupervised neural network based feature extraction using weak top-down constraints

Herman Kamper, Micha Elsner, Aren Jansen, Sharon Goldwater

ICASSP 2015accepted

Deep neural networks (DNNs) have become a standard component in supervised ASR, used in both data-driven feature extraction and acoustic modelling. Supervision is typically obtained from a forced alignment that provides phone class targets, requiring transcriptions and pronunciations. We propose a n…

Cited by 0 Source ↗Scholar ↗