<div dir="ltr"><div>Hello Everyone,<br></div><div>Due to Veterans day our<span style="background-color:rgb(255,229,153)"><b> reading group will be shifted to tomorrow Tuesday at 11</b></span> am in the MAT conference room.<br>
We will have a guest speaker, PhD candidate Maha Alabduljalil, who will present a talk on pairs similarity search. Maha is interested in Similarity search, duplicate detection algorithms, parallelism, and performance optimizations techniques. She has papers in SIGIR, WSDM, and WWW. <br>
Please join us, there will also be special snacks.<br><br><div><b>Talk Abstract:</b></div><div><span style="white-space:pre-wrap"> </span>All
pairs similarity search is used in many applications such as duplicate
detections, data cleaning and clustering. It is still a time consuming
process and in this proposal I present our contributions to speeding up
the process using exact and efficient parallel executions. These
techniques can be summarized as data partitioning, load balancing and
memory optimizations. They have shown to add an order of magnitude over
state of the art algorithms for similarity search using five datasets.</div><div><br></div><div><b>Keywords:</b> <br>Modeling, algorithms, performance</div><br><br></div><div><div><br>-- <br>Saiph Savage<br><br><br>
</div></div></div>