id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_gcwpgxioqfbr7bhenvqbhu2phy	Hoang Cuong	Adapting to All Domains at Once: Rewarding Domain Invariance in SMT	2016	15	.pdf	application/pdf	8676	1121	69	Existing work on domain adaptation for statistical machine translation has consistently assumed access to a small sample from the test Mismatch in phrase translation distributions between test data (target domain) and train data is estimated on out-of-domain training data to a target translation rule with source and target phrases having two similar distributions over the latent subdomains is likely safer to use. , K} encoding (arbitrary) K latent subdomains that generate each source and target phrase ẽ Table 3: Adaptation results when tuning on the in-domain development set. Table 4: Adaptation results when tuning on the mixed-domain development set. • Favouring the source-target coherence across subdomains (i.e., adding the feature D(ẽ, f̃)) provides a significant translation improvement of This paper aims at adapting machine translation systems to all domains at once by favoring phrases that Latent domain phrase-based models for adaptation. models for translation domain adaptation. domain adaptation for statistical machine translation.	./cache/work_gcwpgxioqfbr7bhenvqbhu2phy.pdf	./txt/work_gcwpgxioqfbr7bhenvqbhu2phy.txt