H-MAPS: Hierarchical Memory-Augmented Proactive Search Assistant for Scientific Literature

2026-05-11

AI Summary

This work addresses the high cognitive load and disruption to reading flow caused by manual retrieval during scientific reading, as well as the limited contextual understanding of existing proactive retrieval systems that rely solely on on-screen text, often failing to capture users’ background knowledge and intent. To overcome these limitations, the authors propose a proactive literature exploration assistant grounded in a three-tier hierarchical memory architecture. The system implicitly infers users’ latent information needs from their reading behaviors, translates them into natural language queries, and performs privacy-preserving neural retrieval locally. By integrating domain expertise with real-time reading context, it delivers personalized, non-intrusive literature recommendations. Empirical evaluation demonstrates that, within identical reading scenarios, the system generates domain-specific queries and retrieves highly relevant yet distinct literature for researchers in NLP and HCI, confirming its personalization capability and practical utility.

Abstract

Scientific reading is an active process that frequently requires consulting external resources, but manual keyword searching interrupts the reading flow and imposes a high cognitive load. Existing proactive information retrieval systems often suffer from context ambiguity, as they rely solely on on-screen text and ignore the reader's specific background and intent. In this demonstration, we present H-MAPS (Hierarchical Memory-Augmented Proactive Search Assistant), a proactive literature exploration assistant that resolves this ambiguity by leveraging a three-layered hierarchical memory. Triggered by implicit reading behaviors, H-MAPS articulates the user's latent information needs into explicit natural language questions and performs neural retrieval entirely on the local device to ensure privacy. We demonstrate H-MAPS using a scenario where two researchers, specializing in NLP and HCI, read the same paper. In response, the system generates profile-specific questions and retrieves distinct literature tailored to each user.

Problem

• scientific reading
• proactive information retrieval
• context ambiguity
• cognitive load
• user intent

Innovation

• hierarchical memory
• proactive retrieval
• implicit reading behavior
• local neural retrieval
• personalized literature exploration