EE CS Colloquium: Enabling NLP, Machine Learning, and Few-Shot Learning using Associative Processing, Avidan Akerib, VP of Associative Computing Business Unit, GSI Technologies

EE CS Colloquium<br><br>Title: Enabling NLP, Machine Learning, and Few-Shot Learning using Associative Processing <br>Speaker: Avidan Akerib, VP of Associative Computing Business Unit, GSI Technologies<br>Date: November 8<br>Time: 4:30pm<br>Location: NEC Auditorium, Gates Computer Science Building Room B3 <br><br>Abstract: <br><br>This presentation details a fully programmable, associative,content-based, compute in-memory architecture that changes the concept ofcomputing from serial data processing--where data is moved back and forthbetween the processor and memory--to massive parallel data processing,compute, and search directly in-place.<br><p>This associative processing unit (APU) can be used in many machinelearning applications, one-shot/few-shot learning, convolutional neuralnetworks, recommender systems and data mining tasks such as prediction,classification, and clustering.</p><p> Additionally, the architecture is well-suited to processing large corporaand can be applied to Question Answering (QA) and various NLP tasks suchas language translation. The architecture can embed long documents andcompute in-place any type of memory network and answer complex questionsin O(1). </p><p><br></p><p>Bio: </p><p>Dr. Avidan Akeribs is VP of GSI Technology's Associative ComputingBusiness Unit. He has over 30 years of experience in parallel computingand In-Place Associative Computing. He has over 25 Granted Patents relatedto parallel and in-memory associative computing. Dr. Akeribs has a PhD inApplied mathematics &amp; Computer Science from the Weismann Instiituteof Science, Israel.</p><p>His specialties are Computational Memory, AssociativeProcessing, Parallel Algorithms, and Machine Learning. </p><p><br></p>

Date: 
Wednesday, November 8, 2017 - 4:30pm to 5:30pm
location: 
NEC Auditorium, Gates Computer Science Building Room B3