Collaborative Science Program (CSP) Funding for ASRU 2013 Conference

The ASRU workshop meets every two years and has a tradition of bringing together
researchers from academia and industry in an intimate and collegial setting to
discuss problems of common interest in automatic speech recognition and
understanding. Topics and focus ASRU 2013 will focus on three important topics in
current machine recognition of speech. 1. Neural networks in ASR - from history
to present successes in applications 2. Building ASR systems with limited
resources - development of systems with restricted or no in-domain resources 3.
ASR in applications - interaction with other sources of information. The topic
for the 4th day will be voted by ASRU participants during paper submission.
Papers Submitted papers are encouraged to focus on the areas above, but papers in
other typical ASRU areas (LVCSR Systems, Language Modeling, Acoustic Modeling,
Decoders & Search, Spoken Language Understanding, Spoken Dialog Systems,
Robustness in ASR, Spoken Document Retrieval, Speech-to-Speech Translation,
Speech Summarization, New Applications of ASR and Speech Signal Processing) will
not be penalized. Invited speakers ASRU will be organized in the form of thematic
days featuring each one keynote talk and invited talk(s) or panels. The following
persons have agreed to serve as days organizers: Applications Day: Johan Shalkwyk
and Pedro Moreno (both Google Inc.) Neural network Day: Nelson Morgan (ICSI
Berkeley), Brian Kinsbury (IBM) Limited Resources Day: Jim Glass (MIT) and Aren
Jansen (JHU HLT Center of Excellence) Relevance for US Navy The workshop
concentrates on mining information from spontaneous speech that is crucial for
obtaining timely and precise information in intelligence scenarios. Defense- and
intelligence-oriented US funding agencies have been supporting research in speech
data mining for many years, with the following recent programs: DARPA SPINE,
GALE, TRANSTAC, BOLT, RATS IARPA: BEST, BABEL ASRUs "limited resources" focus is
related to the current IARPA BABEL program concentrating on rapid development of
speech recognition and keyword spotting for under-resourced languages on very
tight schedules.