The active learning prototype is live!

From: https://activelearner.fastforwardlabs.com/

Very cool depth of field rendering experiments by Matt DesLauriers.

From: https://twitter.com/mattdesl/status/1110603724043493376

The Caltech dataset is running in the prototype now.

Quickdraw data and sprites are working in the prototype.

Prototype work continues.

Active learning cover illustration GIF.

Doing more interface work on the active learning prototype. The number sprites are back.

The aesthetic and animation of Petz.

From: https://twitter.com/desandro/status/1104067352730177537

More FF10 cover refinements.

Working on the cover for our next report "Learning with Limited Labelled Data". Planning to generate the illustration from a forked version of @_cjwallace's active learning demo notebook.

Back to work on the prototype. Got the transitions working: select points, label points, retrain model.

I'm working on putting my personal online presence in order. First up, a page for my résumé.

From: https://resume.grantcuster.com/

Using face-tracking and parallax to add depth to a website. Cool idea by Rupert Parry.

From: https://depth-index-demo.glitch.me/

John Caserta's website contains a link to print the website as a book.

From: http://johncaserta.com/

Trying to decide how to show the selected points. This is a little harsh.

This is great and it's going to take me a while to finish. It shows how a thoughtful interface and framing can change how a process feels (compare this to going through your following list in the normal Twitter interface).

From: https://tokimeki-unfollow.glitch.me

I got the MNIST images rendering in the active learning prototype. White here are the images selected for labeling for the next round.

Rasmus Andersson's favorite technical papers.

From: https://www.dropbox.com/sh/is0sy5350lr4v9j/AADQlhVSQcRw6vCNKQgGWelqa

Lots of neat generative sketches by Kjetil Golid.

From: https://generated.space/

First pass of active learning rounds on MNIST in the prototype. I need to think through what the transitions should be more, but this looks promising.