When we think of driving in general, there are good drivers and bad drivers. So, on a 2D spectrum, we would picture a cluster of data of those drivers and realise that the good drivers’ data is clustered around a particular coordinate(x,y) while the bad drivers’ data is all over the place.
On July 16, the ECJ invalidated the EU-US Privacy Shield, one of the key mechanisms for lawfully transferring data between the EU and the US. Federated Learning and other Privacy Preserving techniques would help solve some of the new challenges organizations face.
We’re very excited to announce the next round of open-source software development grants in the OpenMined community, generously sponsored by the University of California San Francisco! These grants will focus on bringing data-centric federated learning with differential privacy budgeting to PyGrid.
We’re very excited to announce the next round of grants sponsored by the PyTorch team! This grant will focus on developing “worker libraries”, allowing PySyft code to be executed in other environments like a mobile phone or web browser.
In medical imaging, necessary privacy concerns limit us from fully maximizing the benefits of AI in our research. These modern privacy techniques could allow us to train our models on encrypted data from multiple institutions, hospitals, and clinics without sharing the patient data.