Search engine rankers powered by pre-trained language models like BERT have shown unprecedented gains in retrieval and ranking tasks, and have quickly become the focus of much of the research in Information Retrieval. Despite the powerful relevance modelling abilities of these methods, there are yet a number of challenges to be solved for these methods to be usable in practice across a wide range of scenarios and search providers; these include, among others: robustness of these rankers to out of distribution data, trade-off between effectiveness, efficiency and hardware, data and training efficiency, effective hybrid sparse-dense methods, integration of feedback methods. In this talk I will give a broad overview of the research done by my research team to address these challenges and I will then deep-dive into some of our methods.



Dr Guido Zuccon
is an Associate Professor in Information Retrieval at the School of ITEE, The University of Queensland; and an ARC DECRA 2018 recipient. Guido also holds honorary positions at Queensland Health as Principal Scientist (Adjunct) and at the University of Strathclyde (UK) as Honorary Reader. Guido’s research in Information Retrieval includes models and methods for search and evaluation, and domain-specific search.



Dr Miao Xu

This session will be conducted in hybrid mode.
UQ St Lucia Campus venue: 46-442 or via Zoom: https://uqz.zoom.us/j/89362232168

About Data Science Seminar

This seminar series will be run as weekly sessions and is hosted by ITEE Data Science.


46-442 or via Zoom