Closing the Training/Inference Gap for Deep Attractor Networks
arXiv:1911.02091
Abstract
This paper improves the deep attractor network (DANet) approach by closing its gap between training and inference. During training, DANet relies on attractors, which are computed from the ground truth separations. As this information is not available at inference time, the attractors have to be estimated, which is typically done by k-means. This results in two mismatches: The first mismatch stems from using classical k-means with Euclidean norm, whereas masks are computed during training using the dot product similarity. By using spherical k-means instead, we can show that we can already improve the performance of DANet. Furthermore, we show that we can fully incorporate k-means clustering into the DANet training. This yields the benefit of having no training/inference gap and consequently results in an scale-invariant signal-to-distortion ratio (SI-SDR) improvement of 1.1dB on the Wall Street Journal corpus (WSJ0).
References in corpus (7)
- Adam: A Method for Stochastic Optimization
- Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separation
- Deep attractor network for single-microphone speaker separation
- Speaker-independent Speech Separation with Deep Attractor Network
- Deep Unfolding: Model-Based Inspiration of Novel Deep Architectures
- Deep Clustering and Conventional Networks for Music Separation: Stronger Together
- FurcaNet: An end-to-end deep gated convolutional, long short-term memory, deep neural networks for single channel speech separation