Learning to Rank with Variable Result Presentation Lengths

2025-06-29

AI Summary

Existing learning-to-rank (LTR) methods assume documents are displayed at fixed lengths, neglecting how presentation format influences user relevance perception and attention allocation. This work introduces *variable-length ranking*—a novel task that jointly optimizes document ranking and vertical space allocation (i.e., display length)—thereby exposing the intrinsic coupling between ordering and presentation. To address this challenge, we propose VLPL, a listwise model built upon the Plackett–Luce framework, equipped with a novel list-level gradient estimation technique for end-to-end optimization. We validate VLPL via semi-synthetic experiments. Results show that VLPL significantly outperforms fixed-length baselines, achieving superior trade-offs between document exposure and attractiveness across diverse scenarios. Even lightweight length-aware mechanisms yield substantial gains, underscoring the importance of explicit length modeling. This work establishes both theoretical foundations and practical paradigms for advancing LTR from “pure ranking” to “ranking + presentation.”

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Search and Optimization: Learning to Search
• Computer Vision: Language and Vision

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Personalized content ranking and presentation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Learning to Rank (LTR) methods generally assume that each document in a top-K ranking is presented in an equal format. However, previous work has shown that users' perceptions of relevance can be changed by varying presentations, i.e., allocating more vertical space to some documents to provide additional textual or image information. Furthermore, presentation length can also redirect attention, as users are more likely to notice longer presentations when scrolling through results. Deciding on the document presentation lengths in a fixed vertical space ranking is an important problem that has not been addressed by existing LTR methods. We address this gap by introducing the variable presentation length ranking task, where simultaneously the ordering of documents and their presentation length is decided. Despite being a generalization of standard ranking, we show that this setting brings significant new challenges: Firstly, the probability ranking principle no longer applies to this setting, and secondly, the problem cannot be divided into separate ordering and length selection tasks. We therefore propose VLPL - a new family of Plackett-Luce list-wise gradient estimation methods for the joint optimization of document ordering and lengths. Our semi-synthetic experiments show that VLPL can effectively balance the expected exposure and attractiveness of all documents, achieving the best performance across different ranking settings. Furthermore, we observe that even simple length-aware methods can achieve significant performance improvements over fixed-length models. Altogether, our theoretical and empirical results highlight the importance and difficulties of combining document presentation with LTR.

Problem

• Optimizing document ranking and presentation lengths jointly
• Addressing user perception changes due to varied document presentations
• Challenging traditional ranking principles with variable length settings

Innovation

• Joint optimization of document ordering and lengths
• Plackett-Luce list-wise gradient estimation methods
• Balances exposure and attractiveness of documents