Rethinking Click Models in Light of Carousel Interfaces: Theory-Based Categorization and Design of Click Models

Jingwei Kang, Maarten de Rijke, Santiago de Leon-Martinez, Harrie Oosterhuis

arxiv.org/abs/2506.18548

Presentation: Friday 18 July 11:55 - ICTIR’25 Oral Presentations

🧵 6/6

Harnessing Pairwise Ranking Prompting Through Sample-Efficient Ranking Distillation

Junru Wu, Le Yan, Zhen Qin, Honglei Zhuang, Paul Suganthan G.C., Tianqi Liu, Zhe Dong, Xuanhui Wang, Harrie Oosterhuis

arxiv.org/abs/2507.04820

Presentation: Thursday 17 July 14:15 - ReNeuIR25 Paper Session

🧵 5/6

Adaptive Orchestration of Modular Generative Information Access Systems

Mohanna Hoveyda, Harrie Oosterhuis, Arjen P. de Vries, Maarten de Rijke, Faegheh Hasibi

arxiv.org/abs/2504.17454

Code: github.com/informagi/AQA

Presentation: Wednesday 16 July 12:15 - Perspectives 1

🧵 4/6

Learning to Rank with Variable Result Presentation Lengths

Norman Knyazev, Harrie Oosterhuis

arxiv.org/abs/2506.23319

Code: github.com/NKNY/varlenr...

Presentation: Wednesday 16 July 10:30 - Reranking

🧵 3/6

Optimizing Compound Retrieval Systems

Harrie Oosterhuis, Rolf Jagerman, Zhen Qin, Xuanhui Wang

arxiv.org/abs/2504.12063

Code: github.com/google-deepm...

Presentation: Tuesday 15 July 11:15 - Search and Ranking 1

🧵 2/6

RecGaze: The First Eye Tracking and User Interaction Dataset for Carousel Interfaces

Santiago de Leon-Martinez, Jingwei Kang, Robert Moro, Maarten de Rijke, Branislav Kveton, Harrie Oosterhuis, Maria Bielikova

arxiv.org/abs/2504.20792

Monday 14 July 14:00 - Resource Papers Posters 1

🧵 1/6

I'm looking forward to reading this, Bhaskar!

Those damn aliens are always so hard to reach! 👽

My colleagues published about pairwise relevance labeling this last year (and even earlier on arXiv):
aclanthology.org/2024.finding...

Why I personally think it works better, mainly because it's hard to calibrate a pointwise relevance prediction, but a pairwise prediction hardly needs calibration.