AI Paper Atlas

Research map / Large language models / Agents, code and math

LLM-Powered Query Expansion for Enhancing Boundary Prediction in Language-Driven Action Localization

Zirui Shang, Xinxiao Wu, Shuo Yang

IJCV 2026 · Large language models › Agents, code and math · approach: Long-horizon LLM agents · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add query-guided temporal modeling (QTM) (ablation)QD-DETR · Charades-STAR1@0.5: 58.17 -> 59.06Improved
add boundary probability modeling (BPM) (ablation)QD-DETR · Charades-STAR1@0.5: 58.17 -> 59.33Improved
add both QTM and BPM (ablation)QD-DETR · TACoSR1@0.7: 20.27 -> 24.04Improved
add QTM (ablation)QD-DETR · TACoSR1@0.5: 38.17 -> 39.37Improved
add BPM (ablation)QD-DETR · TACoSR1@0.5: 38.17 -> 40.04Improved
integrate query expansion and boundary modeling into CG-DETRCG-DETR · Charades-STAR1@0.5: 57.71 -> 58.87Improved
integrate query expansion and boundary modeling into EatrEatr · TACoSR1@0.5: 31.69 -> 34.12Improved
swap LLM for query expansion (Qwen instead of LLaMa3-8B) (ablation)QD-DETR+Ours · Charades-STAR1@0.5: 59.71 -> 59.51No clear effect
randomly swap start and end expanded queries (noisy queries) (ablation)QD-DETR+Ours · Charades-STAR1@0.5: 59.71 -> 59.23No clear effect
let the LLM freely describe objects (Prompt-Object) (ablation)QD-DETR+Ours · TACoSR1@0.5 and R1@0.7: 40.83 -> 39.42; 24.04 -> 21.92Hurt
limit sentences to 10 words (Prompt-Concise) (ablation)QD-DETR+Ours · TACoSR1@0.5: 40.83 -> 40.61No clear effect
use distance plus similarity with expanded query instead of Gaussian smoothing (ablation)QD-DETR+Ours · TACoSR1@0.5: 39.07 -> 40.83Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy