Research map / Video / Action recognition
Compound Memory Networks for Few-shot Video Classification
Linchao Zhu, Yi Yang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add constituent key memory (matrix keys) instead of single-vector key (ablation) | Plain CMN · Kinetics 5-way 1-shot | accuracy: 57.3 -> 60.5 | Improved |
| add constituent key memory (matrix keys) instead of single-vector key (ablation) | Plain CMN · Kinetics 5-way 5-shot | accuracy: 76.0 -> 78.9 | Improved |
| use multi-saliency embedding instead of LSTM embedding (ablation) | LSTM-emb · Kinetics 5-way 1-shot | accuracy: 57.6 -> 60.5 | Improved |
| use CMN instead of Matching Networks | MatchingNet · Kinetics 5-way 5-shot | accuracy: 74.6 -> 78.9 | Improved |
| fine-tune ResNet-50 on meta-training classes before nearest neighbour (ablation) | Nearest-pretrain · Kinetics 5-way 1-shot | accuracy: 51.1 -> 48.2 | Hurt |
| increase memory size from 64 to 512 slots (ablation) | CMN · Kinetics 5-way 1-shot (meta-val) | accuracy: 52.0 -> 55.1 | Improved |
| increase memory size from 512 to 2048 slots (ablation) | CMN · Kinetics 5-way 1-shot (meta-val) | accuracy: 55.1 -> 55.0 | No clear effect |
| use 5 multi-saliency descriptors instead of 1 (ablation) | CMN · Kinetics 5-way 1-shot (meta-val) | accuracy: 53.7 -> 55.1 | Improved |
| increase descriptors from 5 to 10 (ablation) | CMN · Kinetics 5-way 1-shot (meta-val) | accuracy: 55.1 -> 53.2 | Hurt |
| increase number of classes per episode from 5 to 8 way (ablation) | CMN · Kinetics 1-shot (meta-val) | accuracy: 55.0 -> 46.0 | Hurt |
| add one more support example per class (1-shot to 2-shot) (ablation) | CMN · Kinetics 5-way | accuracy: 60.5 -> 70.0 | Improved |
Closest papers by meaning
- Label Independent Memory for Semi-Supervised Few-shot Video Classification (TPAMI 2020 · similarity 0.90)
- Few-Shot Object Recognition From Machine-Labeled Web Images (CVPR 2017 · similarity 0.87)
- Memory Matching Networks for One-Shot Image Recognition (CVPR 2018 · similarity 0.87)
- Generalized Few-Shot Video Classification With Video Retrieval and Feature Generation (TPAMI 2021 · similarity 0.87)
- A Baseline for Few-Shot Image Classification (ICLR 2020 · similarity 0.86)
- Temporal Saliency Query Network for Efficient Video Recognition (ECCV 2022 · similarity 0.86)
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.
