Filter
Exclude
Time range
-
Near
The paper “Split Happens” by Tanveer Khan, Mindaugas Budzys and Antonis Michalas examines leakage in split learning and introduces an FSS-based protocol to keep data and labels on the client. 📰harpocrates-project.eu/split… #SplitLearning #FunctionSecretSharing #MachineLearning
2
27
#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

1
1
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
6
5
29
ML models memorize training data, a major #privacy problem for data collaboration, even w/ #federatedlearning @gg159136 and @vectorinversion gave a fantastic preso at @opendp_org this morning on applying #differentialprivacy to #splitlearning to solve this More to come!
5
13
102
Our new preprint - *FedSL*, we propose a novel Federated and #SplitLearning technique for training sequentially distributed data using #RNN, and show its works better than conventional methods on #EHR data. Great Work @abediai #FederatedLearning #CloudAI tinyurl.com/yxfogn8z

5
Great work being led by @raskarmit @johnkwerner on possible way of using #splitlearning for #covid_19 tracking
1
3
and some of the ongoing MIT projects was really informative and enlightening. Hearing him talk about split learning and federated learning really did open our thinking horizons! (2) #REDX #splitlearning #federatedlearning #innovation #machinelearning #deeplearning #makermela
1
27 Apr 2019
AI on Siloed Data: Sharing wisdom not raw data. Beyond anonymization to protect data, towards #datautility. Specific reference to applications in legal. #ImaginationInAction splitlearning.github.io #Splitlearning #datacommons @jamesondempsey
4
9
Great idea to use #splitlearning & #federatedlearning to enable #ML at a large scale while protecting individual data privacy. Huge implications for #healthcare #imaginationinaction @raskarmit
2
5
.@proneat from @MIT talked about #SplitLearning, a recently developed, highly resource efficient method for collaboration in the health sector, allowing to perform distributed deep learning under data sharing constraints. Download slides bit.ly/2VRy9Uu #datacouncil
1
3
Data Council SF '19 Speaker Announcement: @proneat from @MIT will talk about "#SplitLearning: A Resource Efficient Distributed #DeepLearning Method without Sensitive Data Sharing". Book your tickets now bit.ly/2UdhGx5 #datacouncil
1
1