Machine & Deep learning Israel highlights: 14-23.12.2018

מאת אורי אליאבייב, 23 בדצמבר 2018

וואו, השבוע היה לנו המון פוסטים מעולים וחשובים בקבוצה. ספרתי בקלות יותר מ-15 פוסטים שכאלה ועשיתי את מרבית המאמצים כדי לארגן לכם את כולם במקום אחד מסודר. כפי שאתם בוודאי יודעים, הרשימה הזו נשלחת גם בניוזלטר שלנו כמידי שבוע (הירשמו כאן!). הניוזלטר המלא ניתן למצוא כאן.

Group's Highlights from last week

1. Noam Cohen from the Technion shared with us a wonderful blog post about "A Signal Processing Perspective For Graph Structured Information ". It got many upvoted so don't miss it.

2. Iyar Lin needed to settle an argument with a coworker so he created an interesting poll at the group which got many votes. If that's not enough he just published a test he maid to check some insights he got from the group.

3. Gidi Shperber shared the second part of this blog post "A different kind of (deep) learning: part 2". We all loved the first one so be sure to check this one also.

4. The one and only, Jonathan Laserson, shared with us a blog post he wrote about this work at Zebra medical. If you need to read only one post make it this one.

5. Rani Horev wrote another great explanation about some new article. This time it was about SlowFast from Facebook. If you haven't check Rani's website until now, please do it now. Rani takes any important article and explains it in a good and simple way. Here you can find another example for a blog post he wrote about an article from Google.

6. Yoel Zeldes, an often visitor at our newsletter, shared his blog post about "The Story of a Bad Train-Test Split ". As always, Yoel made a great job so give it a read.

7. Yan Yanko Kotliarsky shared the big news from Facebook "Facebook open-sources its main NLP modelling framework PyText".

8. I have shared a two important blog post from Google and Netflix.

9. Nir Ben-Zvi shared a great NeurIPS 2018 summary for all the NLP people out there.

10. Jenia Golbstein made us all love him thanks for sharing his code with us (SOTA segmentation using Keras).

11. Jacob Gildenblat shared his insights about "Unsupervised Feature Learning via Non-Parametric Instance-level Discrimination". If you need to read only one post, make it this one (#2).

13. Shir Meir Lador told us that the call for proposals of WiDS has just started *and* shared with us a new episode of "Unsupervised". What else can we ask for?

14. Matty Mariansky was brave enough to ask this question.

15. Alex Kazakov invited us to probably one of the best Deep learning events next year.