@techreport{TD:101091,
	att_abstract={The study presented in this work is a first effort at real-time speech translation of TED talks, a compendium of public talks with different speakers addressing a variety of topics. We address the goal of achieving a system that balances translation accuracy and latency. In order to improve ASR performance for our diverse data set, adaptation techniques such as constrained model adaptation and vocal tract length normalization are found to be useful. In order to improve machine translation (MT) performance, techniques that could be employed in real-time such as monotonic and partial translation retention are found to be of use. We also experiment with inserting text segmenters of various types between ASR and MT in a series of real-time translation experiments. Among other results, our experiments demonstrate that a good segmentation is useful, and a novel conjunction-based segmentation strategy improves translation quality nearly as much as other strategies such as comma-based segmentation. It was also found to be important to synchronize various pipeline components in order to minimize latency.},
	att_authors={vk947h, sb7658, jc582p, al1649, rc832b},
	att_categories={},
	att_copyright={},
	att_copyright_notice={},
	att_donotupload={},
	att_private={false},
	att_projects={},
	att_tags={},
	att_techdoc={true},
	att_techdoc_key={TD:101091},
	att_url={http://web1.research.att.com:81/techdocs_downloads/TD:101091_DS1_2013-01-30T21:24:05.497Z.pdf},
	author={Vivek kumar Rangarajan and John Chen and Srinivas Bangalore and Andrej Ljolje and Rathinavelu Chengalvarayan},
	institution={{NAACL-HLT}},
	month={June},
	title={{Segmentation Strategies for Streaming Speech Translation}},
	year=2013,
}