#SplitLearning: Distributed #DeepLearning method without sensitive data sharing. SplitNN is a distributed and private deep learning technique to train deep #neuralnetworks over multiple data sources without the need to share raw labelled data directly. bit.ly/3yiFRNx

How reliably does #splitlearning safeguard private information? @gg159136 and @vectorinversion in our @LiveRamp privacy tech group published their paper on #splitnn attacks and thwarting them with #differentialprivacy in @RealAAAI Check it out: arxiv.org/abs/2201.04018

Split learning is a powerful approach for machine learning #datacollaboration @gg159136 and @vectorinversion on our privacy-enhancing tech team at @LiveRamp participated in a summer fellowship with @opendp_org to combine #differentialprivacy w/ #splitnn liveramp.com/developers/blog…

Is dataset-as-a-service a thing? You start out with a small dataset of your own and pay someone for using their dataset for fine-tuning your model. This is all done without the exchange of data, only model parameters like federated learning or SplitNN

SplitNN-driven Vertical Partitioning deepai.org/publication/split… by Iker Ceballos et al. #DeepLearning #ComputerScience

#splitNN, developed at @MIT in 2018, is a new architectural approach for #ML - opening new roads for innovation - in order to secure the #privacy of #data, Split #NeuralNetworks on #PySyft by Adam James Hall link.medium.com/hQVYhx49h3 #federatedLearning