Now, who got the Friends reference? That's right. It's all about some insights into our peers. With our previously introduced multi-user labeling workflow, it's now time to get some insights in the form of a brand new graph—the inter-annotator agreement. And as if this wasn't enough, we overhauled our embedding creation and active learning module.

TL;DR:
The integration of Websockets makes page refreshing obsolete, the changes from concurrent users are now displayed almost instantly as they happen.
We complete the data slice selection by adding "static" data slices which will always return the same records after the slice was created. An example of when to use this feature can be the static slicing of data in order to distribute them among the labeling workforce.
There are now options to order the records in the data browser. You can order by attribute value, weak supervision confidence and finally you can also order randomly with a customizable seed. Use this feature for data exploration and for better customizable labeling sessions.

In this release, we add the functionality to save filters that was developed in the Data Browser as Data Slices and bring you the first step towards a fully-fledged multi-user integration. Many of you wished for these features for quite some time, and we finally introduce them this patch!

In this release we add a notification center and drastically increase the performance of our application by introducing some upload limits and making changes to the tokenization process.

After a successful first closed beta, we are going in for round 2. And we bring a whole lot of new features.