Select a calendar:
Filter April Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Events for April 12, 2010
-
Sparse Representation Methods for Speech and Language Processing
Mon, Apr 12, 2010 @ 10:00 AM - 11:00 AM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Abstract:
Sparse representation techniques, such as Support Vector Machines (SVMs), k-nearest neighbor (kNN) and Bayesian Compressive Sensing (BCS), can be used to characterize a test sample from a few support training samples in a dictionary set. Traditional Compressing Sensing based methods have been used for signal reconstruction and compression. They have also been successfully applied to the classification of fMRI images. This talk presents our recent work on sparse representations for phonetic classification, speech recognition and text classification in general. The importance of a prior, the sparseness constraint, and choice of the dictionary to this framework will be discussed. Representing a test example as a linear combination of features from the training set allows for a PER on the well-studied TIMIT phone recognition task of 19.0%, which is the best number reported in the literature to-date. Motivated by this result, we also propose a set of features that are a function of the phonetic labels of the original dictionary can be used to create a new representation of the test sample, where the test sample is better linked to the actual units/labels to be recognized.
(Joint work with Tara Sainath, Dimitri Kanevsky and David Nahamoo.)Bio:
Dr. Bhuvana Ramabhadran is the Manager of the Speech Transcription and Synthesis Research Group at the IBM T.J. Watson Center. Upon joining IBM in 1995, she made significant contributions to the ViaVoice line of products focusing on acoustic modeling. She has served as the Principal Investigator of two major international projects: the NSF-sponsored MALACH project, developing algorithms for transcription of elderly, accented speech from Holocaust survivors, and the EUsponsored TC-STAR project, developing algorithms for recognition of EU parliamentary speeches. She has served as the technical chair on conferences, organized workshops, and currently serves on the Speech and Language Technical Committee of the IEEE SPS society. Her research interests include speech recognition algorithms, statistical signal processing, pattern recognition and biomedical engineering.Host: Professor Shrikanth NarayananLocation: Ronald Tutor Hall of Engineering (RTH) - 320
Audiences: Everyone Is Invited
Contact: Mary Francis
-
Recent Development in Simultaneous Recognition of 2 or More Speakers Recorded on a Single Microphone
Mon, Apr 12, 2010 @ 11:00 AM - 12:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Abstract:
Human listeners are able to selectively tune out all but the speaker of interest when multiple speakers are talking, as in a cocktail-party scenario. We describe a new computational method which, for the first time, can accurately transcribe overlapping speakers better than humans can. This is a breakthrough: it is almost unprecedented for computers to outperform people in any field of perception, audio or visual.Bio:
Peder is currently a manager for the speech engines and algorithms group at IBM's TJ Watson Research Center. He is also serving as a chair of the User Interface Technology Professional Interest Community at IBM. Peder has been doing speech recognition at IBM since 1996 and has published over 50 papers in the area. His work in speech has been very mathematical in nature, as is reflected by an Erdos number (http://www.oakland.edu/enp/) of 3 earned while at IBM. Prior to joining IBM, Peder received a PhD in mathematics from the University of Michigan. There he proved that the existence of our universe is reasonable by showing that the Schrodinger equation is well posed and has a unique solution. Presently Peder lives in Westchester, New York, where he is trying to dissuade 6 foot long snakes from taking up residence under his porch.Host: Professor Shrikanth NarayananLocation: Ronald Tutor Hall of Engineering (RTH) - 320
Audiences: Everyone Is Invited
Contact: Mary Francis
-
BME 533 Seminar
Mon, Apr 12, 2010 @ 12:30 PM - 01:50 PM
Alfred E. Mann Department of Biomedical Engineering
Conferences, Lectures, & Seminars
Helga van Herle, M.D., and David Cesario, M.D., Division of Cardiovascular Medicine, USC: Seminar Title: Biomedical Engineering in Cardiology
Location: Olin Hall of Engineering (OHE) - 132
Audiences: Graduate//Department Only
Contact: Mischalgrace Diasanta
-
Exploiting Collective Behavior in Datacenter Storage
Mon, Apr 12, 2010 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
CENG SeminarMithuna Thottethodi, Assistant Professor, Purdue UniversityMonday, April 12, 2010
2:00 – 3:00 p.m.
(RTH -306)Hosted by Prof. Timothy M. PinkstonAbstract:
The storage layer is a significant factor in the capital cost, energy cost and performance of servers and data-centers. Emerging high-performance, low-energy, non-volatile, solid-state storage media show enormous promise to solve two of the three problems (storage performance and energy). However, the high cost-per-byte of solid-state media has hindered wide-spread adoption in servers.
SieveStore is a cost-effective, ensemble-level disk-cache architecture which enables the use of solid-state media to significantly filter access to storage ensembles via two key innovations. First, Sievestore leverages the typical collective behavior of storage ensembles which is quite distinct from the behavior of individual servers in the ensemble.
Second, we show that selective cache allocation – sieving – is fundamental to enable efficient ensemble-level disk-caching. We find that the two components (sieving and ensemble-level caching) each contribute to SieveStore's cost-effectiveness. SieveStore is effective in achieving significantly higher hit ratios than unsieved ensemble-level disk-caching (35%-50% more, on average) while using only 1/7th the number of SSD drives as unsieved, ensemble-level disk-caches.
Further, ensemble-level caching is a strictly superior cost-performance point compared to per-server caching.Biography:
Mithuna Thottethodi is an assistant professor with the School of Electrical and Computer Engineering at Purdue University. His research interests include multicore memory hierarchies and interconnection networks, microarchitecture, datacenters and large storage systems, and programmable microfluidics. He was awarded the National Science Foundation's CAREER award in 2007. He received the Eta Kappa Nu (HKN Beta Chapter) Outstanding Professor award in Fall 2004 and the Ruth And Joel Spira Outstanding Teacher award in 2006. He received his Ph.D. in 2002 from Duke University.Location: Ronald Tutor Hall of Engineering (RTH) - -306
Audiences: Everyone Is Invited
Contact: Janice Thompson
-
Service Orientation: Software Engineering issues beyond Architecture
Mon, Apr 12, 2010 @ 02:00 PM - 03:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Dr. Srinivas Padmanabhuni, Principal Research Scientist, Infosys SETLabsMonday, April 12
2:00 – 3:00 p.m. (HED-116)Hosted by Prof. Viktor K. Prasanna Abstract: Service Oriented Architecture (SOA) is an emergent software architectural trend predicated upon the notion of "service", as a first class software unit for software construction. In this talk, we shall explore the upcoming area of Service Oriented Software Engineering (SOSE), covering software engineering research issues for the different life cycle stages of SOSE based software construction: Service identification, Service based design, Service Contract based Development, Service Based Testing, Service Evolution and Change Management. We shall cover a overview of the diverse Software Engineering issues therein, we shall present our research on the following areas in depth:
1. SOSE process: We present briefly the common approaches of top-down, middle out and bottom up approaches
to SOSE, our focus shall be to present the commonly prevalent business process driven approach to service lifecycle. The key advantage of such an approach is the right sizing and business facing services we shall be able to derive and work with as part of the SOSE approach.
2. Service Identification: A key aspect of service orientation in our approach as above is its close relation to
Business Process Architectures, and we present a homegrown approach to harvesting services from Business processes.
3. Composite Applications: While SOA advocates service based design and reuse at design level, a key practical
requirement in building real SOA based systems is in being able to efficiently combine reuasable services, with custom software, giving rise to new software application trend termed as Composite Applications.
4. Service Evolution and Change Management: We present a plethora of challenges in Service governance and
maintenance, and cover some typical and proprietary approaches for service change management.
5. Service Mining: A key aspect of SOA is the possibilities it offers for legacy systems extension. Key challenges there involve mining of services from legacy portfoliosBiography: Dr. Srinivas Padmanabhuni is a Principal Researcher at Software Engineering and Technology Labs (SETLabs), the R and D arm of Infosys Technologies Limited, Bangalore, India. He supervises the Software Engineering and Distributed Computing research labs in SETLabs at Infosys. Dr. Srinivas specializes in Web services, Service Oriented Architecture, Business Process Management, and Grid technologies alongside pursuing interests in semantic web, autonomic computing, intelligent agents, and enterprise architecture. He has been selected for Who's Who in Asia 2007 first edition, in addition to being nominated for Who is Who in the World and Americas 2009 editions. He serves on editorial board of journals and program committees for international conferences in area of web services, SOA, and Software Engineering. He has authored several papers in international conferences including AAAI, ICWS, SCC, GITMA,APSEC,ISEC, and others. He is currently the chairperson of ACM bangalore chapter. He has served on program committees for several international conferences and workshops including ICWS (International Conference of Web Services), PricAI (Pacific Rim International Conference on AI), NWeSP (International Conference on Next Generation Web Services Practices), Indian Conference on Software Engineering (ICSE) etc. He has given numerous invited speeches at varied industry and academic forums including IEEE and ACM forums, industry CIO seminars like MINDEF CIO Summit, conferences including APSEC, ICWA etc. and architect forums like IASA. He has also authored books and book chapters and articles in some leading professional journals. Prior to Infosys, Dr. Srinivas has worked in multiple capacities in startups out of Canada and USA. Dr. Srinivas holds a doctorate degree in computing science from University of Alberta, Edmonton, Canada. Prior to Ph.D he secured his B.Tech and M.Tech in computer science from Indian Institutes of Technology at Kanpur and Mumbai respectively.Location: Hedco Pertroleum and Chemical Engineering Building (HED) - -116
Audiences: Everyone Is Invited
Contact: Janice Thompson
