dataScienceRetreat

Ups, I meant to add a gif:

This is what happens with a @dataScienceRet mentor when brought in at the right time. You will never know how long you would have struggled with the problem that just got solved. #ml

I remember those times... not in a fond way :)

ML training compute has been doubling every 6 months since 2010 nitter.net/ohlennart/status/… And in the deep learning era, doubling time is 6 months, beating Moore's law! #ml #largedatasets

Less than 3 hours to go... :) Join us live on our youtube channel (link given below) at 6pm CET today (Dec 21st). 1. Estimating the solar potential of buildings using roof Analysis. 2. Ana's Ping Pong 3. Meeting Notes AI 4. Digital Nose 5. Salsa-App

Demo day for batch 27 will begin streaming in 2 hrs. Come over if you are curious about generative models applied to dancing and programming synths. Two tasks that take humans years to master and where novel contributions are rare. meetup.com/Data-Science-Retr… Behind the scene prep:

We are streaming  Demoday for DSR batch26 on youtube.com/channel/UCbj8ZlZ… 18:10 Deep Martin (text simplification) 18:30 Products from Customers' view 18:50 Identifying Disaster Tweets 19:10 Detecting humans in LIDAR point clouds 19:30 Recommender Systems: Two-Tower Deep Neural Network

#beautiful Kinda random for the weekend :)

People with misophonia react to ordinary sounds like loud chewing or smacking with anxiety, rage or even panic. A sound filtering device would make their lives significantly easier. We made an ML tool for automatic detection and location of theses sounds: mtr.cool/menxeyenhm

Have you wondered where someone you crossed on the street bought their clothes? You know can take a pic and you will get where to buy the item. What if the recommendation was from a sustainable source, or second hand? If you like @vinted, this is vinted++ youtube.com/watch?v=P5Yxo6qA…