|Zeit:||7. November 2023, 17:30 Uhr|
|The talk will take place online on WebEx (link below).|
|Download als iCal:||
The Stuttgart ELLIS Unit is pleased to announce the upcoming ELLIS Distinguished Lecture Series talk by Jonathan Berant (Tel-Aviv University / Google DeepMind)!
Title: Large language models and the age of long texts
The recently demonstrated impressive capabilities of pretrained language
models has whet the appetite of many, and generated wide interest in models that can consume and generate long texts. This interest has raised many questions around modeling of long documents and document collections, as well as around their proper evaluation. In this talk, I will discuss developments in this area, including evaluation of models that consume long texts and the role of retrieval-augmented models in this area. In particular I will characterize some of the difficulties that arise when naively augmenting a language models with retrieved information, and present some recent work on joint training of retrievers and language models.
Jonathan Berant is an associate professor at the School of Computer Science at Tel Aviv University, currently on sabbatical as a visiting faculty researcher at Google DeepMind. Jonathan earned a Ph.D. in Computer Science at Tel-Aviv University, and was a post-doctoral fellow at Stanford University, and subsequently a post-doctoral fellow at Google Research, Mountain View. Jonathan has worked in many applied areas of natural language understanding, including semantic parsing, question answering, and textual entailment. Jonathan Received several awards and fellowships including The Rothschild fellowship, The ACL 2011 best student paper award, EMNLP 2014 best paper award, NAACL 2019 best resource paper award, and several honorable mentions. Jonathan has won the Kadar prize for outstanding research and is currently an ERC grantee.