Research map / Optimization / Optimizers
Sampling Wisely: Deep Image Embedding by Top-K Precision Optimization
Jing Lu, Chaofan Xu, Wei Zhang, Ling-Yu Duan, Tao Mei
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 |
|---|---|---|---|
| use top-k precision sampling loss l_k instead of distance weighted sampling | DenseNet201 embedding · CUB-200-2011 | Precision@1 (%): 57.56 -> 61.07 | Improved |
| use l_k instead of semi-hard triplet loss | DenseNet201 embedding · Stanford Cars | Precision@1 (%): 75.16 -> 80.03 | Improved |
| use l_k instead of angular loss | DenseNet201 embedding · Stanford Cars | Precision@1 (%): 77.44 -> 80.03 | Improved |
| use l_k instead of hard mining triplet loss | DenseNet201 embedding · Online Products | Precision@1 (%): 74.98 -> 78.40 | Improved |
| use l_k instead of uniform triplet loss | DenseNet201 embedding · Online Products | Precision@1 (%): 62.89 -> 78.40 | Improved |
| choose negatives by uniform sampling rather than any smarter sampling | triplet loss · all three datasets | Precision@1 | Hurt |
| swap between hard mining, semi-hard and distance-weighted sampling | triplet loss · CUB, Cars, Online Products | Precision@1 | Mixed |
Closest papers by meaning
- Sampling Matters in Deep Embedding Learning (ICCV 2017 · similarity 0.88)
- End-to-End Learning of Deep Visual Representations for Image Retrieval (IJCV 2017 · similarity 0.86)
- Training CNNs With Normalized Kernels (AAAI 2018 · similarity 0.86)
- DAS: Densely-Anchored Sampling for Deep Metric Learning (ECCV 2022 · similarity 0.85)
- Loss Functions for Top-k Error: Analysis and Insights (CVPR 2016 · similarity 0.85)
- Cross-Image-Attention for Conditional Embeddings in Deep Metric Learning (CVPR 2023 · 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.
