A Multi-Perspective Learning to Rank Approach to Support Children's Information Seeking in the Classroom

TL;DR

This paper presents a multi-perspective learning-to-rank model that improves search result relevance for children in classroom settings by considering educational alignment, readability, and appropriateness, enhancing their information discovery experience.

Contribution

It introduces a novel re-ranking model extending listwise learning-to-rank to incorporate multiple classroom-relevant perspectives, improving search results for children.

Findings

The model effectively prioritizes educationally aligned and appropriate resources.

Experimental results outperform existing baselines.

Ablation studies confirm the importance of multiple perspectives.

Abstract

We introduce a novel re-ranking model that aims to augment the functionality of standard search engines to support classroom search activities for children (ages 6 to 11). This model extends the known listwise learning-to-rank framework by balancing risk and reward. Doing so enables the model to prioritize Web resources of high educational alignment, appropriateness, and adequate readability by analyzing the URLs, snippets, and page titles of Web resources retrieved by a given mainstream search engine. Experimental results, including an ablation study and comparisons with existing baselines, showcase the correctness of the proposed model. The outcomes of this work demonstrate the value of considering multiple perspectives inherent to the classroom setting, e.g., educational alignment, readability, and objectionability, when applied to the design of algorithms that can better support children's information discovery.