Data Science Seminar "Text and What Next? AI Across Diverse Data Types"

Language models, such as Le Chat, EuroLLM, Llama, and ChatGPT, have turned the world of artificial intelligence on its head and are increasingly influencing the rest of the world. But they mostly process and generate only text.

The seminar "Text and What Next? AI Across Diverse Data Types" explored AI applications beyond purely textual data. For example: how successfully can AI automatically recognise and produce sign language? How far have we come in understanding speech? Could visual input, combined with language comprehension help robots navigate and behave in ways that make them genuinely useful? The event was led by Mark Fišel, Professor of Natural Language Processing at the University of Tartu Institute of Computer Science.

Speakers:

• Amit Moryossef, Chief Technology Officer at sign.mt, "No Barriers, No Limits: The Future of Sign Language AI"

• Joonas Kalda, PhD student at TalTech, "Towards understanding multi-talker speech"

• Joonas Puura, Senior Machine Learning Engineer at Bolt, "Lessons from Automating Customer Support with LLMs at Bolt"

• Risto Hinno, Data analyst and software developer at Feelingstream, "Lost in transcription"

• Agnes Luhtaru, Visiting research scholar at the University of Texas at Austin, "Robots Facing an Ambiguous World"

Recording of the seminar

Photo gallery of the seminar

The Data Science Seminar is supported by the Ministry of Education and Research Centres of Excellence grant TK213 (Estonian Centre of Excellence in Artificial Intelligence (EXAI)).