User Preferences for Automated Curation of Snackable Content full strip notes
summary
Two user studies compare human edits with automated cutters using classifiers and simple heuristics on movie and TV content: study one shows 150 clips to 1226 responses and finds all four automated cuts (identity, seen, shot, tag) rate higher than the human control; study two shows 96 clips to 580 responses in paired choices and ranks preferences face-based cuts > shot > transcript > control, with face-boundary heuristics most preferred. Findings feed a proof-of-concept platform for automated snackable clips.
AI-written summary · about
related talks
12mTalkLess: Blending Extractive and Abstractive Speech Summarization for Editing Speech to Preserve...
13mCineVision: An Interactive Pre-visualization Storyboard System for Director–Cinematographer Colla...
49mCreation as Curation [Keynote]
33m"Are You Sure?": blending product comparisons and recommendations with A.I.
14mWhy this creator doesn't use AI to make Content
29mQ&A #5
05mMy Little Robot: User Preferences in Game Agent Customization
12mWhy am I seeing this: Democratizing End User Auditing for Online Content Recommendations