AI Paper Atlas

Research map / Video / Video segmentation and tracking

Multiple Pairwise Ranking Networks for Personalized Video Summarization

Yassir Saquil, Da Chen, Yuan He, Chuan Li, Yong-Liang Yang

ICCV 2021 · Video › Video segmentation and tracking · approach: Self-supervised video representation learning · 27 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
train DR-DSN for more epochs (60 -> 2000)DR-DSN · TVSumKendall tau: 0.0169 -> 0.1516Improved
train DR-DSN for more epochs (60 -> 2000)DR-DSN · SumMeKendall tau: 0.0433 -> -0.0159Hurt
use Multi-ranker with personalized preference set {1} instead of Standard rankerMulti-ranker · FineGym personalized, preference set {1}Kendall tau: 0.0254 -> 0.1086Improved
use Multi-ranker with preference set {4} instead of Standard rankerMulti-ranker · FineGym personalized, preference set {4}Kendall tau: 0.1504 -> 0.3007Improved
use Multi-ranker with all preferences (global summary) instead of Standard rankerMulti-ranker · FineGym personalized, preference set {1,2,3,4}Kendall tau: 0.3792 -> 0.3928Improved
use Multi-ranker4 (k-means preferences) instead of Standard ranker for global summarizationMulti-ranker · TVSumKendall tau: 0.1758 -> 0.1736No clear effect
use Multi-ranker4 instead of Standard ranker for global summarizationMulti-ranker · FineGym (50 videos, global summary)Kendall tau: 0.3792 -> 0.3928Improved
use Multi-ranker4 instead of Standard ranker for global summarizationMulti-ranker · SumMeKendall tau: 0.0108 -> -0.0006No clear effect
vary number of k-means preferences (2, 4, 8) (ablation)Multi-ranker · TVSumKendall tau: 0.1630 (2) vs 0.1736 (4) vs 0.1750 (8)No clear effect
vary mini-batch size B (32, 128) and number of pairs N (2000, 5000) (ablation)Standard ranker · TVSum, SumMe, FineGym validationKendall tauNo clear effect
train Standard ranker longer (beyond early epochs) (ablation)Standard ranker · TVSum, SumMe, FineGym validationKendall tauHurt
increase lambda emphasis on global loss (0.25, 0.5, 0.75) (ablation)Multi-ranker · FineGym validation, global Kendall tauKendall tauNo clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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