US vs. UK speaker accent detector made low-latency via deep reinforcement learning.
Seq2Seq-based style transfer between Quebec and Metropolitan French text.
Speech corpora of many languages, styles, and formats exist in the world, representing significant potential for the phonetic sciences. However in practice there are significant practical and methodological barriers to conducting the “same study” …
Semi-supervised detection of rhyme and rhyme schemes, including internal and slant rhyme, using bidirectional LSTMs.
Three models for classification of US vs. UK speaker accent given audio recordings, using several varieties of RNNs and CNNs.
Integration of Kaldi's nnet2 deep neural networks into the Montreal Forced Aligner, permitting the automatic forced alignment of speech audio and a corresponding transcript in time using a Deep Neural Network-Hidden Markov Model architecture.
Predicts dialect of English used to write a SMS-style text message using RNNs supplemented with data from Google Trends.