200 likes | 406 Views
Lending a Hand: Sign Language Machine Translation. Sara Morrissey NCLT Seminar Series 21 st June 2006. Overview. Introduction What, why, how…? Out with the old… SL Corpora The System …in with the new *new and improved* Lost in Translation Evaluation issues Conclusion.
E N D
Lending a Hand: Sign Language Machine Translation Sara Morrissey NCLT Seminar Series 21st June 2006
Overview • Introduction • What, why, how…? • Out with the old… • SL Corpora • The System • …in with the new • *new and improved* • Lost in Translation • Evaluation issues • Conclusion
Introduction • WHAT ? • Sign Language • Visually articulated language • Linguistic phenomena prevalent to SLs • Classifiers • Non-manual features (NMFs) • Discourse mapping and use of signing space
Introduction (2) • WHY ? • a)Improve communication b) Stretching application of EBMT • HOW? • Our approach • Annotated SL corpora • Example-based MT employing Marker Hypothesis (Green, 1979)
Introduction (3) • Other approaches • Transfer - Grieve-Smith, 1999; Marshall & Sáfár, 2002, Sáfár & Marshall 2002; Van Zijl & Barker, 2003 • Interlingua – Veale et al., 1998; Zhao et al., 2000 • Multi-path – Huenerfauth, 2004, 2005 • Statistical – Bauer et al., 1999, Bungeroth & Ney, 2004, 2005, 2006
Out with the old…Corpora • Difficult to find • ECHO project • Nederlandse Gebarentaal (NGT) corpora • 40 minutes of video data • 5 Aesop’s fables by two signers and SL poetry • Combined corpus of 561 sentences
Out with the old… Annotation • Why annotate? • No formal written form for SLs • Linguistic description including NMFs • Can include translation making corpus bi/trilingual • Time for chunking and aligning present • ELAN annotation toolkit • Graphical user interface displaying videos and annotations simultaneously (Fig. 1) • Time-aligned and non-time-aligned annotations including NMF description, repetition notation and notes on indexing and role.
Out with the old…The System • Segmentation using the ‘Marker Hypothesis’ (MH) (Green, 1979) • Analagous to system of (Way & Gough, 2003; Gough & Way, 24a/b) • Segments spoken language sentences according to a set of closed class words • Chunks start with closed class words and usually encapsulate a concept or an attribute of a concept forming concept chunks, e.g <CONJ> or with tiny curls
Out with the old…The System (2) • MH not suitable for use with SL side of corpus due to sparseness of closed class item markers • NGT gloss tier segmented based on time spans of its annotations, remaining annotations with same time span grouped with gloss tier segments forming concept chunks similar to English marker chunks • Despite different methods, they are successful in forming potentially alignable concept chunks
English chunk <CONJ> or with tiny curls NGT chunk <CHUNK> (Gloss RH) TINY CURLS (Gloss LH) TINY CURLS (Repetition RH) u (Repetition LH) u (Eye Gaze) l,d Out with the old…The System (3)
Out with the old…The System (4) • Searches for exact sentence match in aligned bilingual corpus • Uses MH to segment input and searches matching or close match chunks in English side of aligned corpus • Looks for individual words in the bilingual lexicon
Out with the old…Experiments • English and Dutch to NGT (Morrissey & Way, 2005) • 100 sentences • Annotations subjective so evaluation difficult, but promising results • NGT to English Dutch • Traditional MT evaluation metrics can be applied (SER, WER, PER, BLEU) • Sparse output and low scores due to lack of closed class lexical items in NGT • Common marker word insertion
Out with the old…Experiments (2) • SER 96% WER 119% • PER 78% BLEU 0 • Example output and reference translation: • mouse promised help • “you see” said the mouse, “I promised to help you”
…in with the new • New Corpus • ~1400 sentences (SunDial and ATIS corpora) • Flight information queries • ISL signed video version • Homespun annotation • With view to end product • New system • OpenLab
Lost in TranslationEvaluation issues • Mainstream evaluation techniques • Exact text matching • No recognition of synonyms, syntactic structure, semantics • SLs no gold standard • Other possible evaluation metrics • Number of content words/number of words in ref translation • Evaluation of syntactic or semantic relations
Conclusions • Basic system • Corpus problems - Larger corpus such as ISL one in creation, more scope for matches, annotations subjective • EBMT caters for some SL linguistic phenomena • Evaluation metrics unsuitable oral <->non-oral translation
Future Work • Adding in NMF information • Manual analysis • Language model to improve output • Suitable evaluation metrics • Review other writing systems for SLs • Avatar…
Thank You Questions?