ChatGPT and other large language models
På RISE Learning Machines Seminar den 9 februari 2023 ger Joakim Nivre, Erik Ylipää, and Olof Mogren från RISE sin presentation: ChatGPT and other large language models. Seminariet är på engelska.
– ChatGPT has been the AI-talk-of-the-town in 2022. In this seminar, we will have a presentation on what we know about the technology behind ChatGPT followed by a panel discussion about ChatGPT and other large language models, with questions from the audience.
Abstract
ChatGPT has been the AI-talk-of-the-town in 2022. It has demonstrated an impressive fluency, and produced examples of dialog that far surpass the state-of-the-art in language models trained for dialog. The system was presented by OpenAI in late November 2022, and in contrast to other recent releases from them, this one was not accompanied by a research paper. In this seminar, we will have a presentation where Olof Mogren goes through what we know about the technology behind ChatGPT. We will then try a new format for Learning Machines: after the introduction, we will have a panel discussion with Joakim Nivre and Erik Ylipää, discussing some of the aspects of ChatGPT and other large language models. How do they work and how are they trained? How have they affected AI and NLP research? What will be the effects in the future? Finally, we will open the floor to questions and discussions from and with the audience.
Om talarna
Joakim Nivre is a research leader in natural language processing at RISE Research Institutes of Sweden, and professor of computational linguistics at Uppsala University. Erik Ylipää is a researcher at RISE with extensive expertise in transformer-based models. Olof Mogren is a research leader within applied artificial intelligence and deep learning at RISE.