Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Zero-Shot Visual Recognition via Bidirectional Latent Embedding
Qian Wang, Ke Chen
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 SLPP bottom-up latent space before LSM (ablation) | BiDiLEL · UCF101 (81/20), IDT, Att | per-class accuracy (%): 21.2 -> 34.2 | Improved |
| add SLPP bottom-up latent space before LSM (ablation) | BiDiLEL · UCF101 (51/50), IDT, WV | per-class accuracy (%): 8.5 -> 15.4 | Improved |
| add SLPP bottom-up latent space before LSM (ablation) | BiDiLEL · CUB-200-2011, GoogLeNet, Att | per-class accuracy (%): 43.5 -> 49.7 | Improved |
| add SLPP bottom-up latent space before LSM (ablation) | BiDiLEL · AwA, GoogLeNet, Att | per-class accuracy (%): 74.2 -> 72.4 | Hurt |
| add SLPP bottom-up latent space before LSM (ablation) | BiDiLEL · AwA, VGG19, Att | per-class accuracy (%): 79.8 -> 79.1 | Hurt |
| use SLPP instead of LDA for the bottom-up stage (ablation) | BiDiLEL · AwA, GoogLeNet, WV | per-class accuracy (%): 51.1 -> 56.1 | Improved |
| use SLPP instead of LDA for the bottom-up stage (ablation) | BiDiLEL · UCF101 (81/20), C3D, WV | per-class accuracy (%): 31.9 -> 38.3 | Improved |
| use SLPP instead of LDA for the bottom-up stage (ablation) | BiDiLEL · CUB-200-2011, GoogLeNet, WV | per-class accuracy (%): 36.7 -> 34.5 | Hurt |
| use SLPP instead of PCA for the bottom-up stage (ablation) | BiDiLEL · UCF101 (81/20), IDT, Att | per-class accuracy (%): 22.9 -> 34.2 | Improved |
| use supervised LPP instead of unsupervised LPP (ablation) | BiDiLEL · UCF101 (81/20), IDT, WV | per-class accuracy (%): 28.4 -> 32.6 | Improved |
| replace LSM with SVR in the top-down stage (ablation) | BiDiLEL (SLPP) · UCF101 (51/50), C3D, Att | per-class accuracy (%): 20.5 -> 23.8 | Improved |
| replace LSM with SVR in the top-down stage (ablation) | BiDiLEL (SLPP) · AwA, VGG19, Att | per-class accuracy (%): 79.1 -> 75.7 | Hurt |
Closest papers by meaning
- Generalized Zero-Shot Recognition Based on Visually Semantic Embedding (CVPR 2019 · similarity 0.92)
- Generative Zero-Shot Learning via Low-Rank Embedded Semantic Dictionary (TPAMI 2018 · similarity 0.91)
- Latent Embeddings for Zero-Shot Classification (CVPR 2016 · similarity 0.91)
- Rethinking Zero-Shot Learning: A Conditional Visual Classification Perspective (ICCV 2019 · similarity 0.91)
- Multi-Task Zero-Shot Action Recognition with Prioritised Data Augmentation (ECCV 2016 · similarity 0.91)
- TPR: Topology-Preserving Reservoirs for Generalized Zero-Shot Learning (NeurIPS 2024 · similarity 0.90)
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.
