Saturday, January 03, 2009

More Compressed Sensing

In preparation for SODA, I checked out the program, marking off talks I wanted to hear, some on Compressed Sensing, and by the process of browsing via osmotic links and slowly morphing topics, I eventually reached the site for the Geometry and Algorithms meeting at the Center for Computational Intractibility. A lot of great talks here, and now, they are on hi-res video. (I havent got the audio to work yet!). Enjoy.

It is always interesting to see the simple CS ideas reach out to new areas from centralized L_0 to L_1, to distributed CS, to devices and hardware, and now to secrecy. Yaron Rachlin and Dror Baron explore whether CS measurements provide secrecy about the underlying signals, and what can be salvaged.

Labels:

5 Comments:

Anonymous Anonymous said...

Interesting!
Here is a shameless plug: Privacy, often confused with secrecy, also turns out to find CS useful. Though in the somewhat opposite direction: CS ideas imply lower bounds for how noisy private answers to queries must be.
http://research.microsoft.com/apps/pubs/default.aspx?id=64343

--kunal

PS: the above paper talks about the "dual" of CS, but of course the results can be translated to the CS setting.

6:53 AM  
Anonymous Anonymous said...

This is cool, thanks for the pointer, I should have searched deeper for CS + Privacy.

-- metoo

5:35 AM  
Anonymous Anonymous said...

Will you be writing something--in an expository, easy to understand way--on CS?

You write well which is why I visit your blog.

3:57 PM  
Anonymous Anonymous said...

Thank you! I want to write about CS, but am frozen wondering should I emphasize the math or the algorithms or the applications. Each is daunting on its own.

-- metoo

3:58 AM  
Anonymous Anonymous said...

I'm Anon 3:

Applications would be most easily accessible for a general audience. But play to your strength. Other people have written about CS too--Terry Tao, Richard Baraniuk-- and I think they mostly emphasized the application side.

8:50 AM  

Post a Comment

<< Home