Deep Learning

Robust Weak Supervision with Variational Auto-Encoders
A VAE model with specifically designed components to perform weak supervision. Compared to existing weak supervision methods, it is considerably more robust to labelling functions design.
PDF Cite HTML
Trans-Encoder: Unsupervised sentence-pair modelling through self- and mutual-distillations
Bootstrapping an unsupervised sentence encoder by self-distilling knowledge between its bi-encoder and cross-encoder forms, enhancing each other iteratively.
Preprint PDF Cite Code Video HTML Blogpost
PARS: Pseudo-Label Aware Robust Sample Selection for Learning with Noisy Labels
We propose a novel pseudo-label aware robust sample selection method for learning with noisy labels that outperforms state-of-the-art especially in presence of high label noise.
Preprint Cite
Just Mix Once: Mixing Samples with Implicit Group Distribution
Recent work has unveiled how average generalization frequently relies on superficial patterns in data. The consequences are brittle …
Preprint PDF Cite
Improving the expressiveness of neural vocoding with non-affine Normalizing Flows
This paper proposes a general enhancement to the Normalizing Flows (NF) used in neural vocoding. As a case study, we improve expressive …
Preprint PDF Cite DOI
Universal Neural Vocoding with Parallel WaveNet
We present a universal neural vocoder based on Parallel WaveNet, with an additional conditioning network called Audio Encoder. Our …
Preprint Cite Poster Slides DOI
Universal Neural Vocoding with Parallel WaveNet