Profile cover photo
Profile photo
Gokberk Cinbis

Gokberk's posts

Post has attachment

Post has shared content
Very excited about taking on a new role as a director of AI research at Apple in addition to my work at CMU as an associate professor in the Machine Learning Department. Lots of truly amazing AI and Machine Learning projects are taking place all over Apple and I am thrilled to be part of it.

We will be building a team of top-notch Machine Learning and Deep Learning researchers, working on challenging R&D projects as well as conducting fundamental research to advance the state-of-the-art in AI! We are looking for both full-time research scientists as well as postdoctoral fellows to join my team. Apply or ping me if you are interested.*USA&pN=0&openJobId=52662972

Post has attachment

A problem with the arXiv preprints is that quite often they are cited instead of the corresponding peer-reviewed publication. For sure, this harms the traditional citation metrics. But, more importantly, it is confusing to have multiple citations to the very same work.

I think in practice, this problem can be reduced if arXiv would export not only the CoRR reference but also the peer-reviewed reference, once a paper reference is added to the article by the author. (arXiv can easily accept custom bibtex entires instead of plain-text references.)

It could also be handy to have a script that can update arXiv references in a given bibtex file, where appropriate.

How about this incentive to improve the quality and speed of reviews?

Each researcher should have a private karma score based on review quality + review speed, evaluated by assoc. editors and area chairs over time. The submission by an author with a high (low) karma score based on his/her reviewer performance should more likely be assigned to reviewers with high (low) karma scores.

Post has shared content
The Muppets explain consciousness. 

Post has attachment
Linus Torvalds is enthusiastic about deep learning and AI: 

"Now, finally, neural networks are starting to really come into their own. I find that very interesting. It’s not an area I work in, and not really something I foresee working on, but it’s still exciting. And unlike the crazy LISP and Prolog language approaches, recurrent neural networks we know work from nature.  "

Post has attachment
Nice things are happening at XRCE (Grenoble), as covered in this MIT Technology Review article. Congrats to +Adrien Gaidon and the whole team!

Our two recently accepted papers:
1) on weakly supervised object detection (TPAMI)

2) on approximate Fisher kernels and non-iid image models (TPAMI)

The code for the latter one is available on github:

Post has shared content
Very nice photos and descriptions from a trip to Turkey by +Marc Levoy 
Pictures from a 2-week trip through Turkey. Shot using a mix of the Nexus 6 smartphone (using my Google team's HDR+ software) and a Sony a7II mirrorless full-frame camera with various zoom and prime lenses.
Turkey, June 2015
125 Photos - View album
Wait while more posts are being loaded