Peekaboo

So I'm still playing around with constrained parametric min-cuts for object segmentation. A major bottleneck of this algorithm is gPb, the global probability of boundary operator from Malik's group. Luckily, there already is a CUDA implementation of gPb out there: Damascene . Damascene provides a command line interface to apply gPb to ppm images. Since I wanted to include it directly with cpmc, I wrote some mex-wrappers for Damascene. And since I would love to see more algorithms done in Python instead of Matlab, I wrote some Python wrappers, too. You can find both, together with the current version of damascene here . You need a working CUDA setup and the acm library to compile it. Set your paths in common.mk and just "make" it. For the matlab wrappers, you also need to adjust the matlab path in "bindings/Makefile" and "make gpb_mex" in that directory. For the python wrappers, you have to compile damascene with "make shared=1...

There was an interesting talk by Jitendra Malik about "Rich Representations for Learning Visual Recognition" and thereafter a panel discussion with Jitendra Malik, Yann LeCun, Geoff Hinton, Tomaso Poggio, Kai Yu, Yoshua Bengio and Andrew Ng. Many "deep" topics were touched but there is one or two ideas that I found the most noteworthy. The first is the idea by Malik to do "hyper supervision". This is his idea of doing the exact opposite than weak supervision: The training examples are labeled very precisely and with lots of extra information. This makes it possible to find more interesting intermediate representations. It also gives the learning algorithm more to work on. In his introduction he said: "Learning object recognition from bounding boxes is like learning language from a list of sentences." If I understand his ideas correctly, he thinks that is its necessary to have additional clues - like 3D information, tracking and time consist...

Here are some of the quotes of this conference that were more on the lighter side and that made me at least chuckle: Josh Tenenbaum about 1970s linguists: "They didn't have any computers these days. At least none that we would now recognize as computer." Geoff Hinton on Josh Tenenbaums talk about "How to grow a mind": "You made the right generalization of Deep Learning from one example." (Hinting at Josh Tenenbaums interest in one-shot-learning) Some more of Geoff Hinton: On optimizing a non-convex likelihood function: "In good neural network fashion we add noise and momentum and hope for the best." On the same topic: "Now we allow the propabilities to add up to 4. I accidentally forgot to normalize them and they were 2. When I normalized them to one, the result got worse. So I went in the other direction." David W. Hogg in the Sam Roweis symposium: "In astronomy, we work at the photon level. We don't have big b...