Deep Neural Networks (DNNs) are a fundamental tool in the modern development of Machine Learning. Beyond the merits of the training algorithms, a great part of DNNs success is due to the inherent properties of their layered architectures, i.e., to the introduced architectural biases. This talk explores recent classes of DNN models in which the majority of connections are untrained, i.e., randomized or more generally fixed according to some specific heuristic.
Limiting the training algorithms to operate on a reduced set of weights implies intriguing features. Among them, the extreme efficiency of the learning processes is undoubtedly a striking advantage with respect to fully trained counterparts. Besides, despite the involved simplifications, randomized neural systems possess remarkable properties both in practice, achieving state-of-the-art results in multiple domains, and theoretically, allowing us to analyze intrinsic properties of neural architectures.
This talk will cover the major aspects regarding Deep Randomized Neural Networks, with a particular focus on Deep Reservoir Computers for time-series and graphs.
Link Zoom: https://uibuniversitat.zoom.us/j/83043347066
