Computer-assisted pronunciation training—Speech synthesis is almost all you need

All You Need Is Mentorship

I find it humbling to confess that most of the truly original ideas that have driven my research group’s agenda over four decades of time have come, not from my own brain, but instead from the minds of my trainees, both graduate students and post-docs. This on its own might explain why I, rather selfishly, have given them long leashes, allowing them to strike out on their own and craft their ow...

All You Need Is Compassion

The paper presents a new deductive rule for verifying response properties under the assumption of compassion (strong fairness) requirements. It improves on previous rules in that the premises of the new rule are all first order. We prove that the rule is sound, and present a constructive completeness proof for the case of finite-state systems. For the general case, we present a sketch of a rela...

CNN Is All You Need

CNNs have been successfully used in audio, image and text classification, analysis and generation [12,17,18], whereas the RNNs with LSTM cells [5,6] have been widely adopted for solving sequence transduction problems such as language modeling and machine translation [19,3,5]. The RNN models typically align the element positions of the input and output sequences to steps in computation time for ...

All you need is love?

Attention is All you Need

The dominant sequence transduction models are based on complex recurrent or convolutional neural networks that include an encoder and a decoder. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely. E...