<div dir="ltr"><br>Hi Everyone!<br>The <span class="">reading</span> <span class="">group</span> will be TODAY at 2 pm in the MAT conference room. <br>Veronika (from Dr. Gilbert's lab) will present:<br><br><b>Title:</b><a href="http://www.seas.upenn.edu/%7Ejshi/papers/nips00_maila_shi.pdf" target="_blank">Learning Segmentation by Random Walks</a>. <br>
By: Marina Meila and Jianbo Shi.<br>
<b>Abstract:</b><br>
We present a new view of image segmentation by pairwise similarities We<br>
interpret the similarities as edge ows in a Markov random walk and study the<br>
eigenvalues and eigenvectors of the walk s transition matrix This interpretation<br>
shows that spectral methods for clustering and segmentation have a probabilistic<br>
foundation In particular we prove that the Normalized Cut method arises<br>
naturally from our framework Finally the framework provides a principled<br>
method for learning the similarity function as a combination of features <div><br><br></div><div>Hope to see you there! <br></div>
Cheers</div>