Friendly Machine Learning For The Web.
ml5.js aims to make machine learning approachable for a broad audience of artists, creative coders, and students.
Testing the library @ml5js with the coding train tutorial, recognizing images from the webcam. What an amazing tool. It's a great start to begin to understand a lot of new concepts, and with #javascript ❤
ml5js.org/
Processing Community Day Copenhagen is in exactly two weeks! I will introduce @p5xjs, @AndreasRef will introduce @ml5js, a bunch of artists and designers will present their projects, and we provide dinner. Sign up for free here: ida.dk/arrangement/ida-it-pr…
ml4a has a brand new demos page! including javascript in-browser demos and all the figures for the book.
a section for @ml5js demos is in the works for early-mid january.
ml4a.github.io/demos/
Finally, this Wednesday morning (ET) I'll be live streaming @thecodingtrain a one hour session for those with some p5 experience using @ml5js and PoseNet. (youtube link coming soon!)
Great talk about data sonification and working with emotional data from @ahandvanish at #eyeo2018 (Also where she announced @ml5js -Friendly Machine Learning for the web.) ht.ly/wLDm30mL1uA
I was thinking of fun and quick machine learning projects for
#codevember and I built this "Wiki Classifier" using @ml5js. This page
pulls Wikipedia images related to the selected topic and tries to
"guess" what they are.
The students in my visual #machinelearning course @makethinkcode had a great idea - to build a feature extractor using @ml5js which determines the user's proximity to the screen to adjust the readability of a variable width font. Here it is in action.
ALT Showing sentiment analysis results: "I love rainbows and unicorns!" gets a high score (0.99) and "When I forget this dot I am so so terribly sad, that is the worst thing ever in the whole miserable universe.
" gets a very low score (0.004)