Train Deep Learning Models Directly on the Edge!Edgifys framework gives any edge unit the ability to run Deep learning and AI training locally, without the need to extract the data to a server or having to pay the network costs of Model building. We call this self training AI.By training AI models at the edge where the data is generated, one can train on the entire data generated, without compressing, or sampling the data in order to make it easily transferable to the cloud/server, and without the need to anonymise it for privacy reasons. This ability to decentralise the training process, and by doing so, allowing to train on entire datasets, will bring forth a potential new level of accuracy to the world of AI. These new obtainable accuracies can generate both high financial rewards in some verticals (retail, intelligent homes, manufacturing), and life-saving solutions in others (autonomous cars, medical).
Label noise presents a real challenge for supervised learning algorithms. Consequently, mitigating label noise has attracted immense research in recent years. Noise robust losses is one of the more promising approaches for dealing with label noise, as these methods only require changing the loss ...
We tackle the problem of Federated Learning in the non i.i.d. case, in which local models drift apart, inhibiting learning. Building on an analogy with Lifelong Learning, we adapt a solution for catastrophic forgetting to Federated Learning. We add a penalty term to the loss function, compelling ...
In The News
Edge computing startup Edgify secures $6.5M Seed from Octopus, Mangrove and semiconductor | TechCrunch
