مشروع ترانزيتيود (Transitude) يركز على دور الترجمة الآلية في نشر معلومات مضللة على وسائل التواصل الاجتماعي، من خلال تحليل تأثير الترجمة الآلية على آراء ومعتقدات طالبي اللجوء واللاجئين حول المجتمعات المضيفة والمواقف الناتجة عن ذلك.

Funded by the @IrishResearch & hosted by Dublin City University @DCU, #Transitude has more to offer. Looking forward to discovering what else is in the #Transitude bag. Stay tuned!. Links to the paper & datasets along with the code will be shared soon.

Exciting time to present #Transitude project in an #Arabic-speaking country, beautiful #Tunisia, which makes this event both professionally rewarding and personally meaningful.

Three publications have emerged from the hashtag#Transitude project so far, exceeding the initial goal of one publication from the first stage (achieving this shows that what seems like over-promising once can, in fact, be possible 😉 ).

This achievement would not have been possible without the contribution of Mohamed Kadhem Jamii, Ilyes Ben Khalifa, Montassar Ben Messaoud, Sondess MISSAOUI & @Jossmo. Special thanks to those involved in organising this conference & to reviewers for providing invaluable reviews.

Novel Findings: It reveals how MT usability is perceived & used in real-world contexts, identifying usage patterns, common translation themes & factors influencing user perception, especially with the introduction of Neural MT & large language models (LLMs) & their impact on MT.

Quality Evaluation Framework for MT: We employ state-of-the-art machine learning and natural language processing techniques, including sentiment analysis and topic modeling, to deliver robust insights into AI-powered MT services from the user perspective.

Major contributions include:. Extensive Dataset: Approximately 4.9 million + 20K user entries from Google Play, X and Reddit. New Benchmark (UEwMT): A large-scale benchmark, constructed from a comprehensive collection of online user-generated content.

Our work addresses this by using a large dataset of user-generated content, consisting of social media posts and app reviews.

With the widespread adoption of MT, there is a real need to understand how users feel about it, their satisfaction levels, and the factors that shape these perceptions. Previously, studies were often limited in scope, resorting to subjective methods.

Our study investigates user feedback on Artificial Intelligence (AI)-powered MT services across various platforms. This is the first comprehensive study, examining MT quality & usability from the user’s perspective, analysing how they perceive & interact with these services.

Our paper, “UEwMT: Leveraging User Experience and Deep Learning for Evaluating Machine Translation Services,” has been accepted for presentation at the Service-Oriented Computing for AI Applications workshop at ICSOC:

RT @ufal_cuni: @OndrejBojar @prajdabre1 @zouharvi @esalesk We continued with great talks by @RicardoRei7, @julius_gulius, @onadegibert, @jo….

يركز المشروع على اللاجئين وطالبي اللجوء الناطقين باللغة العربية الذين انتقلوا حديثًا إلى أيرلندا. سندعم ذلك بأدلة مبنية على البيانات التي تظهر مدى جودة الترجمة الآلية وتأثيرها على مستخدميها.

Transitude also participated in ADAPT's 3rd Annual Scientific Conference last May. Great opportunity to share our project, network, and explore future collaborations! Thanks to @IrishResearch for funding Transitude, @DCU for hosting @KhetamAlSharou & @Jossmo for his support.

Such a pleasure presenting at @EAMT_2024! The discussions were enriching, & the overwhelmingly positive feedback has been truly satisfying. Transitude was described as innovative, practical, and inspiring, to mention just a few. Proud of the project & what we have achieved so far

Happy to announce that our paper "Transitude: Machine Translation on Social Media: MT as a potential tool for opinion (mis)formation" has been accepted for presentation as part of the Projects/Products track of the EAMT conference in Sheffield, June 24-28. @KhetamAlSharou @Jossmo.