Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Energy Confused Adversarial Metric Learning for Zero-Shot Image Retrieval and Clustering
Read the paperCheck 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 |
|---|---|---|---|
| Add Energy Confusion regularization to Binomial metric learning | Binomial on CARS · CARS | R@1: 74.2 -> 84.5 | Helped |
| Add Energy Confusion regularization to Binomial metric learning | Binomial on CUB · CUB | R@1: 52.9 -> 55.7 | Helped |
| Add Energy Confusion regularization to Binomial metric learning | Binomial on Stanford Online Products · Stanford Online Products | R@1: 68.5 -> 71.3 | Helped |
| Add Energy Confusion regularization to Binomial metric learning | Binomial on In-Shop · In-Shop | R@1: 81.7 -> 83.8 | Helped |
| Add Energy Confusion regularization to Triplet metric learning | Triplet on CARS · CARS | R@1: 68.3 -> 81.0 | Helped |
| Add Energy Confusion regularization to N-Pair metric learning | N-Pair on Stanford Online Products · Stanford Online Products | R@1: 68.0 -> 69.8 | Helped |
| Set λ to 0.02 instead of 0 (ablation) | ECAML(Tri) on CARS · CARS | R@1: 68.3 -> 81.0 | Helped |
| Set λ to 0.3 instead of 0 (ablation) | ECAML(N-Pair) on CARS · CARS | R@1: 74.3 -> 80.4 | Helped |
| Increase λ beyond 0.02 to 1 (ablation) | ECAML(Tri) on CARS · CARS | R@1: 81.0 -> 59.3 | Hurt |
| Increase embedding size from 128 to 512 (ablation) | ECAML(Binomial) on CARS · CARS | R@1: 79.6 -> 84.5 | Helped |
| Increase embedding size from 128 to 512 (ablation) | Binomial on CARS · CARS | R@1: 70.8 -> 74.2 | Helped |
| Apply dropout with ratio 0.25 (ablation) | Binomial on CARS · CARS | R@1: 74.2 -> 74.5 | No clear effect |
Closest papers by meaning
- Hybrid-Attention Based Decoupled Metric Learning for Zero-Shot Image Retrieval (CVPR 2019 · similarity 0.89)
- Divide and Conquer the Embedding Space for Metric Learning (CVPR 2019 · similarity 0.88)
- Signal-To-Noise Ratio: A Robust Distance Metric for Deep Metric Learning (CVPR 2019 · similarity 0.88)
- Deep Factorized Metric Learning (CVPR 2023 · similarity 0.87)
- Deep Metric Learning via Lifted Structured Feature Embedding (CVPR 2016 · similarity 0.87)
- Learning Low-Dimensional Metrics (NeurIPS 2017 · similarity 0.87)
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.