Andrea Santilli, PhD student in AI at GLADIA, Sapienza University of Rome and former PiSchool alumnus, is now working on Natural Language Processing, Representation Learning and Reasoning, focusing on investigating language models’ zero-shot generalization and reasoning abilities.
In this talk, Andrea will present the collaborative work done at Huggingface’s BigScience initiative, a one-year-long research workshop on large multilingual models and datasets.
The proposed model T0 outperformed GPT-3 on 9/11 tasks despite being 16x smaller.
This event is only one of the Tech Talks that Pi School of AI organize during the year to spread the word about the latest AI technologies and applications.
The Tech Talk will be live on 01/12/22 at 4:00 PM GMT on Google Meet.
Register in order to receive an invitation.