Private Distributed Learning in a Byzantine World
Georgios Damaskinos - Machine Learning Engineer at Facebook- spoke at the ML Fairness Summit in June 2021.
The ever-growing number of edge devices - e.g., smartphones - and the exploding volume of sensitive data they produce, call for distributed machine learning techniques that are privacy-preserving. Given the increasing computing capabilities of modern edge devices, these techniques can be realized by pushing the sensitive-data-dependent tasks of machine learning to the edge devices and thus avoid disclosing sensitive data.
I will present two important challenges in this new computing paradigm along with an overview of our proposed solutions to address them. First, for many applications, such as news recommenders, data needs to be processed fast, before it becomes obsolete. Second, given the large number of uncontrolled edge devices, some of them may undergo arbitrary -Byzantine - failures and deviate from the distributed learning protocol with potentially negative consequences such as learning divergence or even biased predictions.
Key Takeaways:
- Our data is extremely valuable and vulnerable = let's push it to the "Edge"
- Machine Learning at the Edge is possible yet challenging due to A. temporality of the data and B. unreliability of the machines
Видео Private Distributed Learning in a Byzantine World канала RE•WORK
The ever-growing number of edge devices - e.g., smartphones - and the exploding volume of sensitive data they produce, call for distributed machine learning techniques that are privacy-preserving. Given the increasing computing capabilities of modern edge devices, these techniques can be realized by pushing the sensitive-data-dependent tasks of machine learning to the edge devices and thus avoid disclosing sensitive data.
I will present two important challenges in this new computing paradigm along with an overview of our proposed solutions to address them. First, for many applications, such as news recommenders, data needs to be processed fast, before it becomes obsolete. Second, given the large number of uncontrolled edge devices, some of them may undergo arbitrary -Byzantine - failures and deviate from the distributed learning protocol with potentially negative consequences such as learning divergence or even biased predictions.
Key Takeaways:
- Our data is extremely valuable and vulnerable = let's push it to the "Edge"
- Machine Learning at the Edge is possible yet challenging due to A. temporality of the data and B. unreliability of the machines
Видео Private Distributed Learning in a Byzantine World канала RE•WORK
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
![Eugenio Culurciello, Founder, TeraDeep - RE.WORK Deep Learning Summit 2015](https://i.ytimg.com/vi/bX-WYlFlx-U/default.jpg)
![Interview with Adam Cheyer, Co-Founder and VP Engineering/VP of R&D at Viv Labs/Samsung](https://i.ytimg.com/vi/el2TTvrb4xM/default.jpg)
![In Search of Trustworthy & Ethical AI | Women in AI Podcast | Rachel Alexander, Omina Technologies](https://i.ytimg.com/vi/IJmm0hc43L0/default.jpg)
![Challenges of Human Labour Automatization in Transport - Juris Puce - Deep Learning Summit #reworkDL](https://i.ytimg.com/vi/Ssj0RG6mrJg/default.jpg)
![Superior Conversational Customer Experience for the Enterprise with Rasa](https://i.ytimg.com/vi/oSuVLNRqfW0/default.jpg)
![ML Observability: A Critical Piece in Ensuring Responsible AI | Arize AI](https://i.ytimg.com/vi/2FE1sg749Vo/default.jpg)
![Training Models at Facebook Scale with PyTorch](https://i.ytimg.com/vi/EXFjhF4HQ_Y/default.jpg)
![Interview with Jeff Stein, Orbital Insight - RE•WORK Connect Summit #reworkAI](https://i.ytimg.com/vi/hNUF6jq430s/default.jpg)
![How Motion-Based, Off-Grid Renewable Energy Will Power the IoT - Jessica O. Matthews, Uncharted Play](https://i.ytimg.com/vi/Xz29ORuQjaE/default.jpg)
![This week is the Deep Learning Summit & Deep Learning in Healthcare Summit Boston](https://i.ytimg.com/vi/htqYmfP6YZ8/default.jpg)
![Wearables: Moving Beyond Sensing - Sonny Vu, CEO, Misfit Wearables](https://i.ytimg.com/vi/Q9G4qw2tlGo/default.jpg)
![7-Eleven Using Applied AI to Disrupt Convenience - Shahmeer Mirza, 7Eleven](https://i.ytimg.com/vi/LC8Dui7MZNQ/default.jpg)
![Interview With Jeremy Marvel, National Institute of Standards and Technology](https://i.ytimg.com/vi/VEuRtXyllFo/default.jpg)
![Elaine Warburton, QuantuMDx - RE.WORK & #BehindTheFace](https://i.ytimg.com/vi/ydvt2HEazYw/default.jpg)
![Rainer Sternfeld, Founder & CEO, Planet OS - IOT Summit, London 2015](https://i.ytimg.com/vi/I18hC7C3NuA/default.jpg)
![The RE.WORK IoT Summit, London, 2015](https://i.ytimg.com/vi/5ApuB6CZn9U/default.jpg)
![Rob May, CEO, Talla - RE•WORK Virtual Assistant Summit, San Francisco 2016 #reworkAI](https://i.ytimg.com/vi/uV7j8N3NZCM/default.jpg)
![Automating Data Annotation with MicroModels with Sama](https://i.ytimg.com/vi/juaX6RH_Siw/default.jpg)
![How Magical Technology will Revolutionize Human Health - Juhan Sonin, Involution Studios](https://i.ytimg.com/vi/nGV0HD4ZAnk/default.jpg)
![IoT Summit, Boston 2015 - with MIT, Xively, INEX Advisors, SafeLogic & more #reworkconnect](https://i.ytimg.com/vi/G0-46Bzs8Zk/default.jpg)
![Growing the Intelligence of Future City Infrastructure, Eric Yeatman, Imperial](https://i.ytimg.com/vi/qKM24mJQ0yE/default.jpg)