Going Beyond T-SNE: Exposing whatlies in Text Embeddings

no code implementations • EMNLP (NLPOSS) 2020 • Vincent Warmerdam, Thomas Kober, Rachael Tatman

We introduce whatlies, an open source toolkit for visually inspecting word and sentence embeddings.

Dimensionality Reduction Sentence +2

Going Beyond T-SNE: Exposing \texttt{whatlies} in Text Embeddings

1 code implementation • 4 Sep 2020 • Vincent D. Warmerdam, Thomas Kober, Rachael Tatman

We introduce whatlies, an open source toolkit for visually inspecting word and sentence embeddings.

Dimensionality Reduction Sentence +2

Non-lexical Features Encode Political Affiliation on Twitter

no code implementations • WS 2017 • Rachael Tatman, Leo Stewart, Am Paullada, alynne, Emma Spiro

Previous work on classifying Twitter users{'} political alignment has mainly focused on lexical and social network features.

Gender and Dialect Bias in YouTube's Automatic Captions

no code implementations • WS 2017 • Rachael Tatman

This project evaluates the accuracy of YouTube{'}s automatically-generated captions across two genders and five dialect groups.

Automatic Speech Recognition Automatic Speech Recognition (ASR) +2

``Oh, I've Heard That Before'': Modelling Own-Dialect Bias After Perceptual Learning by Weighting Training Data

no code implementations • WS 2017 • Rachael Tatman

Human listeners are able to quickly and robustly adapt to new accents and do so by using information about speaker{'}s identities.

Automatic Speech Recognition Automatic Speech Recognition (ASR) +1