disco icon indicating copy to clipboard operation
disco copied to clipboard

Decentralized & federated privacy-preserving ML training, using p2p networking, in JS

Results 128 disco issues
Sort by recently updated
recently updated
newest added

**New Fair and Robust aggregation function server-side** Described here: https://arxiv.org/abs/2012.04221 We discussed with Felix and it seems like it is one of the most SOA aggregation function to add to...

feature

Currently in the tasks.json, the description is html code, so when we load it we are then injecting it into the vue template. It is bad practice to inject html...

priority
rework
web client

Using console.log can be very slow (similar to how print() in python or other languages console output can create a big overhead). This is important only in the training_manager as...

feature
for later
dev env

as an example use-case, let's build a decentralized model for browser-based images we can start from a stand-alone-version based on https://github.com/justadudewhohacks/face-api.js/ (uses TF.js) see for example https://www.codeproject.com/Articles/5276827/AI-Age-Estimation-in-the-Browser-using-face-api-an on top of...

for later

For each task there is only one "event" that people can join, it might be interesting to be able to have multiple "events" per task to join. A (silly) example...

for later

label label / label images is confusing for now, and we should add a bit more info what example dataset people can use / download where to start playing with...

good first issue

cifar 10 or 100. or even go to imagenet directly :) in the task description we can maybe include a link how people can download an (arbitrary) part of this...

We are currently using all-reduce scheme for model averaging. We should add the option to use RelaySGD. I will start working on this.

We have to look into how to allow devices behind NAT to send and receive p2p messages. Options: - LibP2P Built-in Mechanisms - https://en.wikipedia.org/wiki/TCP_hole_punching

For tabular datasets (popular examples: [adult income](https://www.kaggle.com/uciml/adult-census-income) and [titanic](https://www.kaggle.com/c/titanic)), normalization is critical for neural network approaches. The most typical and a very effective way to normalize is to "subtract the...

feature