ASR2000 - Automatic Speech Recognition: Challenges for the new Millenium September 18-20, 2000 Paris, France

Breaking the Local Optima Paradigm: DARPA Speech Research Initiatives in Multi-Modal and Other Technologies

James D. Bass

Defense Advanced Research Projects Agency (DARPA), Information Technology Office, Arlington, VA, USA

Direct government funding for ASR research has been on a steady decline since 1994. DARPA, The National Science Foundation (NSF), and other agencies have reduced funding in anticipation of product development critical mass from the private sector. While private industry has conducted research and delivered several speech related products to the marketplace, specific military requirements for robust speech recognition in adverse environments are not being actively addressed by private industry. After presenting these views to the directors of various government research organizations, new projects are in the works to enhance the reliability and adaptive nature of ASR systems. The purpose of this paper is to inform the international community of these new initiatives in hopes of expanding the research base.

Full Paper (PDF)   Full Paper (Zipped Postscript)

Bibliographic reference.  Bass, James D. (2000): "Breaking the local optima paradigm: DARPA speech research initiatives in multi-modal and other technologies", In ASR-2000, 241-243.