Linguistically Enhanced Text to Sign Gloss Machine Translation

In spite of the recent advances in Machine Translation (MT) for spoken languages, translation between and Sign Languages (SLs) or remains a difficult problem. Here, we study how Neural (NMT) might overcome communication barriers Deaf Hard-of-Hearing (DHH) community. Namely, approach Text2Gloss task which text segments are translated to lexical sign representations. this context, leverage transformer-based models via (1) injecting linguistic features that can guide learning process towards better translations; (2) applying Transfer Learning strategy reuse knowledge pre-trained model. To aim, different aggregation strategies compared evaluated under random weight initialization conditions. The results research reveal successfully contribute achieve more accurate models; meanwhile, procedure applied conducted substantial performance increases.