Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
A Boundary Based Out-of-Distribution Classifier for Generalized Zero-Shot Learning
Xingyu Chen, Xuguang Lan, Fuchun Sun, Nanning Zheng
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 the classification loss L_cls to the SVAE objective (ablation) | boundary OOD classifier · AWA1 | AUC: 62.5 -> 94.9 | Improved |
| add the classification loss L_cls to the SVAE objective (ablation) | boundary OOD classifier · CUB | AUC: 56.1 -> 98.2 | Improved |
| add the cross-reconstruction loss L_cr to the SVAE objective (ablation) | boundary OOD classifier · AWA1 | AUC: 62.5 -> 89.3 | Improved |
| add the cross-reconstruction loss L_cr to the SVAE objective (ablation) | boundary OOD classifier · CUB | AUC: 56.1 -> 60.6 | Improved |
| add L_cr on top of L_cls (full objective) (ablation) | boundary OOD classifier · AWA1 | AUC: 94.9 -> 96.8 | Improved |
| add L_cr on top of L_cls (full objective) (ablation) | boundary OOD classifier · CUB | FPR: 9.2 -> 1.1 | Improved |
| increase latent dimension from 64 to 128 (ablation) | boundary OOD classifier · AWA1 | AUC: 96.8 -> 90.5 | Hurt |
| increase latent dimension from 64 to 256 (ablation) | boundary OOD classifier · CUB | AUC: 99.6 -> 95.1 | Hurt |
| reduce latent dimension from 64 to 16 (ablation) | boundary OOD classifier · AWA1 | AUC: 96.8 -> 95.2 | Hurt |
| replace confidence-based soft gating (CB-GATING-3) with boundary-based hard OOD gate | gating for GZSL · CUB | FPR at 95% TPR: 72.0 -> 2.5 | Improved |
| replace CB-GATING-3 with boundary-based OOD gate | gating for GZSL · AWA1 | FPR at 95% TPR: 45.5 -> 12.5 | Improved |
| gate f-CLSWGAN with the boundary OOD classifier instead of using it alone | f-CLSWGAN · AWA1 GZSL | harmonic mean H: 59.6 -> 72.6 | Improved |
Closest papers by meaning
- Distinguishing Unseen From Seen for Generalized Zero-Shot Learning (CVPR 2022 · similarity 0.92)
- A Review of Generalized Zero-Shot Learning Methods (TPAMI 2022 · similarity 0.91)
- An Empirical Study and Analysis of Generalized Zero-Shot Learning for Object Recognition in the Wild (ECCV 2016 · similarity 0.91)
- Improving Generalized Zero-Shot Learning by Exploring the Diverse Semantics from External Class Names (CVPR 2024 · similarity 0.91)
- Data Distribution Distilled Generative Model for Generalized Zero-Shot Recognition (AAAI 2024 · similarity 0.90)
- Data-Free Generalized Zero-Shot Learning (AAAI 2024 · similarity 0.89)
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.
