Profile cover photo
Profile photo
Pratik Chaudhari
629 followers
629 followers
About
Communities and Collections
View all
Posts

Post has attachment
Old San Juan
Old San Juan
photos.google.com
Add a comment...

Post has shared content
Pi isn’t the only thing we’re celebrating today. Find out how #AlphaGo discovered a beautiful move that no one saw coming: http://goo.gl/WNHY6E via +WIRED
Photo
Add a comment...

Post has shared content

Post has shared content
SynthCam3D

http://robotvault.bitbucket.org/

Could I draw your kind attention to our new library of annotated synthetic indoor scenes we have been working on for a while? We think this could be potentially very useful for people interested in scene understanding for robotics and generating unlimited amount of training data from arbitrary view-points. This library emerged as a part of our joint-reconstruction-and-semantic-segmentation with conv-nets based approach. We have been trying to do segmentation of functional categories of objects purely from geometric cues.  

Our semantic segmentation module builds on the work which +Vijay Badrinarayanan kindly involved me in last year using his idea of saving pooling indices inspired from Marc'Aurelio Ranzato's unsupervised learning method. It is quite essential when do you semantic segmentation that you get your boundaries right and saving pooling indices in your conv-net whenever you use pooling helps quite a lot.  You can find the relevant papers at the bottom of the webpage and nice little demo  +Alex Kendall  created together with +Kesar Breen  's caffe implementation. 

The library webpage also has a small presentation in the publications section that I gave this CVPR in a workshop organised by Ian Reid. This is joint work with +Vijay Badrinarayanan, +Viorica Patraucean, +Simon Stent  and Roberto Cipolla and we are happy to share all the models and hope that you can use them and give us good feedback on expanding the library and the overall approach. 
Add a comment...

Post has attachment
Add a comment...

Post has attachment
AI-Stats 2015, San Diego
Trip to San Diego
Trip to San Diego
plus.google.com
Add a comment...

Post has shared content

Post has shared content
The paper linked to below is chosen fairly arbitrarily from Neural Information Processing Systems, an open access journal with open peer review. The advantages of open peer review are clear: the reader gets useful extra information about a paper by reading what expert reviewers think about it. What about the disadvantages? I have heard it suggested that authors may be humiliated in public, or that reviewers will not be willing to share their thoughts publicly. But the few examples I've seen of open peer review show no evidence of this kind of problem. For example, if you click on "Reviews" in the page below, you are taken to a page where there are three anonymous reviews and responses by the authors. What stands out for me is just how polite and constructive the discussion is. It makes it clear that open review certainly can work very well. Maybe it sometimes leads to problems, but I've never seen any, so I think those risks are greatly outweighed by the advantages.
Add a comment...

Post has shared content

Post has shared content
Yesterday 47 years ago MLK was killed, and RFK delivered what may well be the greatest, most moving impromptu speech ever. I don't really know how he could possibly prepare for such a difficult speech (the audience, mostly black, hadn't heard the news yet) in just 30 minutes. Yet it easily ranks among the very best speeches ever given. Simply amazing.
Add a comment...
Wait while more posts are being loaded