We provide novel guaranteed approaches for training feedforward neural networks with sparse connectivity. We leverage on the techniques developed previously for learning linear networks and show that they can also be effectively adopted to learn non-linear networks. We operate on the moments involving label and the score function of the input, and show that their factorization provably yields t...

My visit at KTH aimed to develop a system in which audio-visual feedback is provided to users who are free to move in the space. The feedback depends on the expressivity of the body movements they perform. Specifically, users can render a music performance by controlling in realtime acoustic parameters through their full-body movement and get a visual feedback on a screen in front of them with ...