Machine translation in support of sign language interpretation and linguistic accessibility – reality or utopia? 

Uncategorized @fi

Päivi Rainò and Özer Çelik

The huge advances in machine translation technology, especially between spoken languages, suggest that similar progress could easily be made between sign and spoken languages. However, applications and project proposals that boldly offer promising solutions and translation services leave those seeking credible evidence empty-handed. Even the most recent international, generously funded research projects leave behind relatively modest research papers and results that were already known in advance (cf. Jantunen et al., 2021; SignOn Project, 2022; Shterionov et al., 2023).  

None of the AI-based machine translation solutions offered so far have been able to take into account the multidimensional and complex structure of sign languages, due to the different information structure of sign language texts compared to, for example, Indo-European or Finno-Ugric languages spoken in Europe (cf. Jantunen et al. 2021; De Weerdt, forthcoming). Moreover, the emphasis on the need for machine translations for sign language users’ access to information has tended to ignore the multilingual nature of sign language users (cf. Rainò, 2020; Jantunen, 2024).  

Despite the above, this presentation will provide examples of why research projects using machine translation technology should be included, despite modest expectations and results. We describe the working processes behind the AI-based translation algorithms currently being developed for Finnish and Finnish Sign Language (FinSL), and the surprising side-stream data that this process has revealed about FinSL. We also consider how the knowledge generated by our project can be used – albeit in small steps – to promote linguistic accessibility also for the deaf people using sign language interpreter services. 

Literature

De Weerdt, D. (2024, tulossa). You sign the way you reason, or theirs? – International Days of Linguistic Accessibility, 16.-17.5.2024, Helsinki. 

Jantunen, T., Rousi, R., Rainò, P., Turunen M., Valipoor, M. & García, N. (2021). Is there any hope for developing automated translation technology for Sign Languages? Teoksessa M. Hämäläinen, N. Partanen & K. Alnajjar (toim.) Multilingual Facilitation. Helsinki: University of Helsinki. https://helda.helsinki.fi/server/api/core/bitstreams/779fe09f-df3b-4446-ae40-9d8d482a5202/content 

Jantunen, T. (2024). Yksikielisyysvinouma, kielikäsitykset ja yhteiskunnallinen kielityö. – Kieli, koulutus ja yhteiskunta 16. [Ilmestyy maaliskuussa 2024.] 

Rainò, P. (2021). Viittomakielibarometri 2020. Tutkimusraportti. Oikeusministeriön julkaisuja, Selvityksiä ja ohjeita 2021:4. Helsinki: Oikeusministeriö. http://urn.fi/URN:ISBN:978-952-259-860-8  

Shterionov, D., De Sisto, M., Vandeghinste, V., Brady, A., De Coster, M., Leeson, L., Blat, J., Picron, F., Scipioni, M. P., Parikh, A., ten Bosh, L., O’Flaherty, J., Dambre, J. & Rijckaert, J. (2022). Sign Language Translation: Ongoing Development, Challenges and Innovations in the SignON Project. In Proceedings of the 23rd Annual Conference of the European Association for Machine Translation, pages 325–326, Ghent, Belgium. European Association for Machine Translation. 

SignON Project (2022). Sign Language Technology: Do’s and Don’ts A Guide to Inclusive Collaboration Among Policymakers, Researchers, and End Users [White paper]. 26th December 2022. https://signon-project.eu/wp-content/uploads/2023/12/White-Paper-Sign-Language-Technology-FINAL.pdf  

Last modified: 20.2.2024