Research map / Video / Video segmentation and tracking
Fast Video Classification via Adaptive Cascading of Deep Models
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use WEG adaptive specialization and cascading instead of always running the oracle | VGGFace face recognition · Friends, Good Will Hunting, Ellen Show, The Departed, Ocean's Eleven / Twelve | GPU latency: 2.4-7.8× | Helped |
| Use WEG adaptive specialization and cascading instead of always running the oracle | VGGFace face recognition · Friends, Good Will Hunting, Ellen Show, The Departed, Ocean's Eleven / Twelve | accuracy | Mixed |
| Disable adaptive window exploration and use a fixed window size of 30 or 60 (ablation) | WEG face recognition · Friends | speedup | Hurt |
| Set training skew to the observed dominant-class skew instead of the fixed training skew (ablation) | WEG face recognition · Friends | accuracy: 30% | Hurt |
| Set training skew to the observed dominant-class skew instead of the fixed training skew (ablation) | WEG face recognition · Friends | speedup | Helped |
| Exit specialization when current skew falls below the skew observed at entry, rather than using estimated accuracy (ablation) | WEG face recognition · Friends | latency: more than 50% | Hurt |
| Use WEG when face-class distributions change every 10, 20, or 30 seconds | WEG face recognition · Synthetic traces with changing distributions | speedup: 0.95/1.19/1.48× | Mixed |
Closest papers by meaning
- Accelerated Training for Massive Classification via Dynamic Class Selection (AAAI 2018 · similarity 0.86)
- AdaFrame: Adaptive Frame Selection for Fast Video Recognition (CVPR 2019 · similarity 0.86)
- LookWhen? Fast Video Recognition by Learning When, Where, and What to Compute (NeurIPS 2026 · similarity 0.86)
- SpeedNet: Learning the Speediness in Videos (CVPR 2020 · similarity 0.86)
- FASTER Recurrent Networks for Efficient Video Classification (AAAI 2020 · similarity 0.86)
- FasterCache: Training-Free Video Diffusion Model Acceleration with High Quality (ICLR 2025 · similarity 0.85)
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.