DiscoSign: Discourse-Aware Text to Sign Language Gloss Translation
Apple's new model tackles discourse-level sign language translation — a long-ignored gap in accessibility AI.
Why it matters
DiscoSign extends LLM-based sign language translation beyond single sentences to discourse-level phenomena (coreference, question-answer structures), advancing accessibility AI from capability research into linguistic grounding.
The key facts
10 to knowApple research: discourse-aware sign language gloss translation
Addresses spatial coreference resolution across discourse
Handles Question-Answer Clauses (QACs) and pseudocleft structures
Modular LLM-based framework grounded in linguistic research
Published Sep 2026 on Apple ML research blog
Apple Research published DiscoSign framework
Addresses three discourse phenomena: spatial coreference resolution, Question-Answer Clauses (QACs), role shifts
Modular LLM-based approach grounded in linguistic research
Focuses on discourse-level translation vs. traditional sentence-level systems
Published September 11, 2026 on Apple ML research blog
Go to the source
Apple Machine Learningmachinelearning.apple.com
Publisher excerpt: Sign language processing systems have traditionally operated at the sentence level, ignoring critical discourse phenomena fundamental to sign language comprehension. We introduce DiscoSign, a computational approach for discourse-aware text to sign language gloss translation grounded in linguistic…