Events Calendar

NL Seminar- Beyond Parallel Data: Joint Word Alignment and Decipherment Improves Machine Translation [EMNLP Practice Talk]

Speaker: Qing Dou, USC/ISI

Talk Title: Beyond Parallel Data: Joint Word Alignment and Decipherment Improves Machine Translation [EMNLP Practice Talk]

Series: Natural Language Seminar

Abstract: Inspired by previous work, where de- cipherment is used to improve machine translation, we propose a new idea to combine word alignment and decipher- ment into a single learning process. We use EM to estimate the model parameters, not only to maximize the probability of parallel corpus, but also the monolingual corpus. We apply our approach to im- prove Malagasy-English machine transla- tion, where only a small amount of paral- lel data is available. In our experiments, we observe gains of 0.9 to 2.1 Bleu over a strong baseline.

Biography: Qing Dou is a fifth year Ph.D. student at ISI. He works with Professor Kevin Knight on various decipherment problems and its application to different Natural Language Processing tasks.

Host: Aliya Deri and Kevin Knight

More Info: http://nlg.isi.edu/nl-seminar/

Location: Information Science Institute (ISI) - Conference Room # 689, Marina del Rey

Audiences: Everyone Is Invited

Contact: Peter Zamar

Event Link: http://nlg.isi.edu/nl-seminar/