Disco greatly simplify access to distributed machine learning, It runs in a browser, allowing it to work on pretty much any platform, simply open the link in a browser and you're using it.
It allows creating your own machine learning tasks so that others can help with theses by just loading their data in the browser. By computing theses tasks, you create a machine learning model, a mathematical system which allows a program to make predictions based on a given set of data. It fits many use cases, such as determining the likeliness that some hospital patient has a disease, or automatically determine what an image represents.

Disco is in full development, adding new features and algorithms regularly. It aims at being the breeding ground for new machine learning technologies. Here goes a list of papers to be integrated into it

• PowerGossip: Practical Low-Rank Communication Compression in Decentralized Deep Learning
• Decentralized Stochastic Optimization and Gossip Algorithms with Compressed Communication
• Practical Secure Aggregation for Privacy-Preserving Machine Learning
• Secure Byzantine-Robust Machine Learning
• Asynchronous Decentralized SGD with Quantized and Local Updates
• Learning from History for Byzantine Robust Optimization
• Byzantine-Robust Learning on Heterogeneous Datasets via Bucketing

---

For more information, contact the C4DT Factory