Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
April 1, 2026Open Access

LSAT Focus: EAGLE’s Embedded Entities Highlighting Technique for NTCIR-18 Lifelog-6

View Full Paper
Ask AI
Bookmark
Share

Authors

TNThang-Long Nguyen-HoATAllie TranMTMinh-Triet Tran

Discussion

Loading...

Member takes

Overview

This work demonstrates improved retrieval accuracy in lifelog moments using embedding and semantic relevance strategies, suggesting better query design.

Key Points

  • The aim is to enhance automatic searching methods for identifying distinct life moments using advanced retrieval strategies.
  • Conducted experiments in the Lifelog Semantic Access Task (LSAT) at NTCIR-18.
  • Compared retrieval strategies, including keyword matching, embedding extraction, and hybrid methods.
  • Focused on directing model attention to key query terms and semantic relevance.
  • The best-performing method improved retrieval accuracy using embeddings with extended descriptions.
  • Hybrid methods showed less effectiveness, likely due to limitations in keyword-matching algorithms.
  • Findings emphasize the need for richer descriptive entities to enhance retrieval outcomes.

Cite This Study

Nguyen-Ho et al. (2025) studied this question.

synapsesocial.com/papers/69cd7b065652765b073a8b24https://doi.org/10.20736/0002002050
View Full Paper
Ask AI
Bookmark
Share