Difference between revisions of "OpenMined"
m |
m |
||
| Line 12: | Line 12: | ||
*** [[Cybersecurity]] | *** [[Cybersecurity]] | ||
* [http://www.openmined.org Openmined] | * [http://www.openmined.org Openmined] | ||
| − | ** [http://github.com/OpenMined/PySyft PySyft] a A library for encrypted, [[ | + | ** [http://github.com/OpenMined/PySyft PySyft] a A library for encrypted, [[privacy]] preserving deep learning |
** [http://github.com/OpenMined/syft.js syft.js] a client-side microlibrary for running PySyft operations in [[Javascript]] | ** [http://github.com/OpenMined/syft.js syft.js] a client-side microlibrary for running PySyft operations in [[Javascript]] | ||
| − | * [http://arxiv.org/abs/1811.04017 A generic framework for | + | * [http://arxiv.org/abs/1811.04017 A generic framework for privacy preserving deep learning | T. Ryffel, A. Trask, M. Dahl, B. Wagner, J. Mancuso, D. Rueckert, & J. Passerat-Palmbach] |
* [[Machine Learning as a Service (MLaaS)]] | * [[Machine Learning as a Service (MLaaS)]] | ||
* [[Watch me Build a Cybersecurity Startup]] | [[Creatives#Siraj Raval|Siraj Raval]] | * [[Watch me Build a Cybersecurity Startup]] | [[Creatives#Siraj Raval|Siraj Raval]] | ||
| Line 24: | Line 24: | ||
** [http://github.com/ethereumjs/testrpc Ethereum Test RPC | GitHub] | ** [http://github.com/ethereumjs/testrpc Ethereum Test RPC | GitHub] | ||
* [http://medium.com/fondercom/a-new-business-model-decentralized-ownership-7b05adac2df4 A New Business Model: Decentralized Ownership | Fonder Open Network - Medium] | * [http://medium.com/fondercom/a-new-business-model-decentralized-ownership-7b05adac2df4 A New Business Model: Decentralized Ownership | Fonder Open Network - Medium] | ||
| − | * [[Privacy in Data Science | + | * [[Privacy]] in Data Science |
Conceptually, federated learning proposes a mechanism to train a high quality centralized model while training data remains distributed over a large number of clients each with unreliable and relatively slow network connections. OpenMined can be considered an implementation of a [[federated]] learning architecture which powers decentralization using blockchain smart contracts...[http://towardsdatascience.com/technology-fridays-openmined-powers-federated-ai-using-the-blockchain-d124e6560dd OpenMined Powers Federated AI Using the Blockchain | Jesus Rodriguez - Towards Data Science] | Conceptually, federated learning proposes a mechanism to train a high quality centralized model while training data remains distributed over a large number of clients each with unreliable and relatively slow network connections. OpenMined can be considered an implementation of a [[federated]] learning architecture which powers decentralization using blockchain smart contracts...[http://towardsdatascience.com/technology-fridays-openmined-powers-federated-ai-using-the-blockchain-d124e6560dd OpenMined Powers Federated AI Using the Blockchain | Jesus Rodriguez - Towards Data Science] | ||
| Line 35: | Line 35: | ||
# Syft — The library containing Neural Networks that can be trained in an encrypted state (so that Miners can’t steal the neural networks that they download to train). | # Syft — The library containing Neural Networks that can be trained in an encrypted state (so that Miners can’t steal the neural networks that they download to train). | ||
| − | OpenMined a platform that merges [http://www.globalsign.com/en/blog/glossary-of-cryptographic-algorithms/ cryptographic] techniques such as [http://en.wikipedia.org/wiki/Homomorphic_encryption Homomorphic Encryption] and [http://en.wikipedia.org/wiki/Secure_multi-party_computation multi-party computation] and [https://en.wikipedia.org/wiki/Blockchain Blockchain] technology to create the ability to train Machine Learning (ML) models with private user data. OpenMined allows AI companies to develop models, have them trained on user data without compromising user [[ | + | OpenMined a platform that merges [http://www.globalsign.com/en/blog/glossary-of-cryptographic-algorithms/ cryptographic] techniques such as [http://en.wikipedia.org/wiki/Homomorphic_encryption Homomorphic Encryption] and [http://en.wikipedia.org/wiki/Secure_multi-party_computation multi-party computation] and [https://en.wikipedia.org/wiki/Blockchain Blockchain] technology to create the ability to train Machine Learning (ML) models with private user data. OpenMined allows AI companies to develop models, have them trained on user data without compromising user [[privacy]], and incentivize users to train their model. OpenMined is a community focused on building technology for [http://blog.openfuture.io/technical/what-is-the-future-for-decentralized-ownership/ decentralized ownership] of data and AI. Data scientists can pay users directly for their data and train AI models in a decentralized way. |
<youtube>HaulMYiB4EQ</youtube> | <youtube>HaulMYiB4EQ</youtube> | ||
Revision as of 22:17, 26 September 2020
Youtube search... ...Google search
- Case Studies
- Openmined
- PySyft a A library for encrypted, privacy preserving deep learning
- syft.js a client-side microlibrary for running PySyft operations in Javascript
- A generic framework for privacy preserving deep learning | T. Ryffel, A. Trask, M. Dahl, B. Wagner, J. Mancuso, D. Rueckert, & J. Passerat-Palmbach
- Machine Learning as a Service (MLaaS)
- Watch me Build a Cybersecurity Startup | Siraj Raval
- Decentralized: Federated & Distributed Learning
- Homomorphic Encryption
- Blockchain Smart Contracts
- Interplanetary File System (IPFS) - a framework that aims to replace the centralized architecture of the HTTP protocol of the internet with a more distributed approach
- Ethereum is a global, open-source platform for decentralized applications
- A New Business Model: Decentralized Ownership | Fonder Open Network - Medium
- Privacy in Data Science
Conceptually, federated learning proposes a mechanism to train a high quality centralized model while training data remains distributed over a large number of clients each with unreliable and relatively slow network connections. OpenMined can be considered an implementation of a federated learning architecture which powers decentralization using blockchain smart contracts...OpenMined Powers Federated AI Using the Blockchain | Jesus Rodriguez - Towards Data Science
Specifically, the OpenMined architecture is based on four fundamental building blocks:
- Capsule — A PGP server to generate public and private keys in order to guarantee the integrity of the different components of a Sonar neural network.
- Sonar — The heart of the OpenMined platform, Sonar is a federated learning server running on the blockchain that control the execution of the different parts of a deep learning appliucation. This library communicates with the Capsule to generate PGP keys and deliver the final, trained results back to the Data Scientist. It also communicates with miners, collecting Gradients and distributing Bounty accordingly.
- Mine — This component hosts the individual data repositories for specific users. Mines regularly interact with Sonar to detect new neural nets to contribute to. The more data that is uploaded to a mine, the more relevant it becomes to Sonar.
- Syft — The library containing Neural Networks that can be trained in an encrypted state (so that Miners can’t steal the neural networks that they download to train).
OpenMined a platform that merges cryptographic techniques such as Homomorphic Encryption and multi-party computation and Blockchain technology to create the ability to train Machine Learning (ML) models with private user data. OpenMined allows AI companies to develop models, have them trained on user data without compromising user privacy, and incentivize users to train their model. OpenMined is a community focused on building technology for decentralized ownership of data and AI. Data scientists can pay users directly for their data and train AI models in a decentralized way.
Contents
Andrew Trask
How Do We Democratize Access to Data?" | Siraj Raval
- OpenMined_demo - code for this video
Building SocketWorker
Programming OpenMined
