Apple Machine Learning Research has introduced DiscoSign, a novel framework for translating text into sign language glosses that accounts for discourse phenomena. This LLM-based approach addresses spatial coreference resolution, question-answer clauses, and concept-gloss consistency, which are crucial for sign language comprehension but often overlooked by sentence-level systems. DiscoSign also introduces new evaluation metrics to assess discourse-level quality, demonstrating significant improvements in spatial consistency and entity tracking compared to traditional methods. AI
IMPACT This research advances AI's capability in sign language translation, potentially improving accessibility for the Deaf and Hard-of-Hearing community.
RANK_REASON The item describes a new computational approach and framework for text to sign language gloss translation published by Apple's research division. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Apple Machine Learning Research →
- American Sign Language
- Apple Inc.
- Apple Machine Learning Research
- ASL STEM Wiki
- DiscoSign
- FLEURS-ASL
- Gallaudet University
- Leah Findlater
- Lorna C Quandt
- Mert Inan
- Northeastern University
- Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
- Raja Kushalnagar
- Vasileios Baltatzis
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →