Rapid development of machine translation for low density languages

1
Rapid development of machine translation for low
density languages

• New Ideas
• Use machine learning to learn translation rules
  from native speakers who are not trained in
  linguistics or computer science.
• Multi-Engine translation architecture can
  flexibly take advantage of whatever resources are
  available.
• Seeded version spaces for learning MT in
  data-poor situations.

Interface for data elicitation

• Impact
• Rapid and low-cost development of machine
  translation for languages with scarce resources.
• OMNIVOROUS MT (from scarce parallel corpora or
  indigenous informants).
• Responsive to rapidly changing intelligence needs.

Schedule
September 01 Seeded Version Space learning
concept prototype (SVS). January 02
Example-Based Machine Translation (EMBT). July
02 Multi-Engine EBMT, statistical MT, and
SVS-ready evaluation of data-poor conditions.
Carnegie Mellon University, Language Technologies
Institute L. Levin, J. Carbonell, A. Lavie, R.
Brown, J. Lafferty