[Google speaker] Attributed Text Generation

Dr. Ni Lao from Google research will talk about their most recent work, released on October 17, 2022. It reveals what it takes to mitigate one of the biggest challenges for large language models.
Talk abstract:
Language models (LMs) now excel at many tasks such as few-shot learning, question answering, reasoning, and dialog. However, they sometimes generate unsupported or misleading content. A user cannot easily determine whether their outputs are trustworthy or not, because most LMs do not have any built-in mechanism for attribution to external evidence. To enable attribution while still preserving all the powerful advantages of recent generation models, we propose RARR (Retrofit Attribution using Research and Revision), a system that 1) automatically finds attribution for the output of any text generation model and 2) postedits the output to fix unsupported content while preserving the original output as much as possible. When applied to the output of several state-of-the-art LMs on a diverse set of generation tasks, we find that RARR significantly improves attribution while otherwise preserving the original input to a much greater degree than previously explored edit models. Furthermore, the implementation of RARR requires only a handful of training examples, a large language model, and standard web search.

Speak Bio:
Dr. Ni Lao is an expert in Machine Learning, Natural Language Understanding and Knowledge graph. He was the Chief scientist and co-founder of Mosaix.ai, a voice AI startup. He is now a tech and research lead at Google research. Ni received his Ph.D. in Computer Science from CMU. [http://cs.cmu.edu/~nlao](http://cs.cmu.edu/~nlao)

Related paper: https://arxiv.org/abs/2210.08726

7-7:05 pm Meet and greet
7:05-7:50pm Presentation
7:50-8:00 Q&A