Smart Compose: Using Neural Networks to Help Write Emails
Google shared some information about their new feature. Most important: they claim to focus on Fairness and Privacy, training on completely anonimized data and trying to eliminate biases.
Link: https://ai.googleblog.com/2018/05/smart-compose-using-neural-networks-to.html
#Google #SmartCompose #FairAI #Privacy
Google shared some information about their new feature. Most important: they claim to focus on Fairness and Privacy, training on completely anonimized data and trying to eliminate biases.
Link: https://ai.googleblog.com/2018/05/smart-compose-using-neural-networks-to.html
#Google #SmartCompose #FairAI #Privacy
Googleblog
Smart Compose: Using Neural Networks to Help Write Emails
Plug-and-play differential privacy for your tensorflow code
#GoogleAI has just released a new library for training machine learning models with (differential) privacy for training data.
where you would write
instead just swap in the
Tutorial: https://github.com/tensorflow/privacy/blob/master/tutorials/mnist_dpsgd_tutorial.py
Link: https://github.com/tensorflow/privacy
#Privacy #tensorflow
#GoogleAI has just released a new library for training machine learning models with (differential) privacy for training data.
where you would write
tf.train.GradientDescentOptimizer
instead just swap in the
DPGradientDescentOptimizer
Tutorial: https://github.com/tensorflow/privacy/blob/master/tutorials/mnist_dpsgd_tutorial.py
Link: https://github.com/tensorflow/privacy
#Privacy #tensorflow
GitHub
privacy/tutorials/mnist_dpsgd_tutorial.py at master Β· tensorflow/privacy
Library for training machine learning models with privacy for training data - tensorflow/privacy
ββTOWARDS FEDERATED LEARNING AT SCALE: SYSTEM DESIGN
Google published how they do #FederatedLearning at scale on tens of millions of mobile phones. This is about training model on decentralized data.
ArXiV: https://arxiv.org/pdf/1902.01046.pdf
#Google #Privacy
Google published how they do #FederatedLearning at scale on tens of millions of mobile phones. This is about training model on decentralized data.
ArXiV: https://arxiv.org/pdf/1902.01046.pdf
#Google #Privacy
Estimating the success of re-identifications in incomplete datasets using generative models
99.98% of Americans would be correctly re-identified in any dataset using 15 demographic attributes, suggesting that even heavily sampled anonymized datasets are unlikely to satisfy the modern standards for anonymization set forth by GDPR.
This is a big concern about privacy and a problem for Data Engineering, especially for those working with anonymized personal information. Paper provides a way to re-identify person from anonymized dataset, this can be useful for people who work for government or security companies
https://www.reddit.com/r/science/comments/chko43/9998_of_americans_would_be_correctly_reidentified/
#privacy #gdpr #federatedlearning #ml
99.98% of Americans would be correctly re-identified in any dataset using 15 demographic attributes, suggesting that even heavily sampled anonymized datasets are unlikely to satisfy the modern standards for anonymization set forth by GDPR.
This is a big concern about privacy and a problem for Data Engineering, especially for those working with anonymized personal information. Paper provides a way to re-identify person from anonymized dataset, this can be useful for people who work for government or security companies
https://www.reddit.com/r/science/comments/chko43/9998_of_americans_would_be_correctly_reidentified/
#privacy #gdpr #federatedlearning #ml
Reddit
From the science community on Reddit: 99.98% of Americans would be correctly re-identified in any dataset using 15 demographicβ¦
Explore this post and more from the science community
Image "Cloaking" for Personal Privacy
New research project from the University of Chicago CS group claims to provide a new face recognition protection mechanism.
Project link: https://sandlab.cs.uchicago.edu/fawkes/
Github: https://github.com/Shawn-Shan/fawkes
#Privacy #DL #CV #facerecognition
New research project from the University of Chicago CS group claims to provide a new face recognition protection mechanism.
Project link: https://sandlab.cs.uchicago.edu/fawkes/
Github: https://github.com/Shawn-Shan/fawkes
#Privacy #DL #CV #facerecognition
GitHub
GitHub - Shawn-Shan/fawkes: Fawkes, privacy preserving tool against facial recognition systems. More info at https://sandlab.cβ¦
Fawkes, privacy preserving tool against facial recognition systems. More info at https://sandlab.cs.uchicago.edu/fawkes - Shawn-Shan/fawkes