Deep Learning for Signal Processing

About this talk

Deep learning is a powerful technique for tasks as classification or recognition. However, it can be used in signal processing for tasks that need to recover the full signal like denoising or source separation. I will present my work on deep learning for music source separation, particularly some strategies for data augmentation which makes the framework more robust to variations one might expect in real-life. This research is reproducible and is accompanied by a python library which makes it easy to work with large datasets.