Monday, 21st November 2022 - 17:00 pm Tatsu Hashimoto : Seminar

Title:   Diffusion based language models

Abstract:

Over the last few years, language models such as GPT-3 have proven to be immensely useful, providing new kinds of few-shot learning and reasoning abilities. However, the discrete, autoregressive nature of these language models leads to challenges when generating long texts, or enforcing constraints on the output. In recent work, we developed diffusion-based language models that provide a promising approach to address these drawbacks and enable new forms of complex control. Despite these promising initial results, many challenges remain in making diffusion language models as powerful and easy to use as their image counterparts, and we will discuss ongoing work to address those challenges

Bio:

Tatsunori Hashimoto is an Assistant Professor in the Computer Science Department at Stanford University. He is a member of the statistical machine learning and natural language processing groups at Stanford, and his research uses tools from statistics to make machine learning systems more robust and reliable — especially in challenging tasks involving natural language generation. His work has received the best paper runner-up at the International Conference on Machine Learning and the best paper at the Neural Information Processing Systems workshop on Networks. Before becoming an Assistant Professor, he was a postdoctoral researcher at Stanford with Percy Liang and John Duchi and received his Ph.D. from MIT under the supervision of Tommi Jaakkola and David Gifford.

Add to your calendar

 vCal  iCal