Research map / Detection and segmentation / Open-set and few-shot detection
Unsupervised Learning of Discriminative Attributes and Visual Representations
Check 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 |
|---|---|---|---|
| learn binary attributes using clustered concepts and discriminative hashing | unsupervised CNN features · CIFAR-10 | mAP: 17.21 vs 16.96 | Helped |
| learn binary attributes using clustered concepts and discriminative hashing | unsupervised CNN features · STL-10 | accuracy: 76.8 ± 0.3 vs 75.4 ± 0.3 | Helped |
| learn binary attributes using clustered concepts and discriminative hashing | unsupervised CNN features · Caltech-101 | accuracy: 89.4 ± 0.5 vs 87.2 ± 0.6 | Helped |
| use patch-wise attribute-code neighbor retrieval | contour detection · BSDS500 | AP: 0.82 vs 0.80 | Helped |
| replace modified discriminative clustering with k-means (ablation) | proposed method · CIFAR-10 | mAP: 17.21 -> 17.16 | Hurt |
| replace modified discriminative clustering with k-means (ablation) | proposed method · STL-10 | accuracy: 76.8 ± 0.3 -> 76.2 ± 0.3 | Hurt |
| replace modified discriminative clustering with k-means (ablation) | proposed method · Caltech-101 | accuracy: 89.4 ± 0.5 -> 88.6 ± 0.5 | Hurt |
| replace modified clustering and CNN feature learning with k-means and HOG features (ablation) | proposed method · CIFAR-10 | mAP: 17.21 -> 16.94 | Hurt |
| replace modified clustering and CNN feature learning with k-means and HOG features (ablation) | proposed method · STL-10 | accuracy: 76.8 ± 0.3 -> 75.5 ± 0.3 | Hurt |
| replace modified clustering and CNN feature learning with k-means and HOG features (ablation) | proposed method · Caltech-101 | accuracy: 89.4 ± 0.5 -> 87.3 ± 0.6 | Hurt |
| replace proposed hashing with Triplet hashing (ablation) | proposed method · CIFAR-10 | mAP: 17.21 -> 17.17 | No clear effect |
| replace proposed hashing with Triplet hashing (ablation) | proposed method · STL-10 | accuracy: 76.8 ± 0.3 -> 76.7 ± 0.4 | No clear effect |
Closest papers by meaning
- Unsupervised Deep Learning by Neighbourhood Discovery (ICML 2019 · similarity 0.90)
- Unsupervised Deep Learning via Affinity Diffusion (AAAI 2020 · similarity 0.89)
- Unsupervised Representation Learning by Predicting Image Rotations (ICLR 2018 · similarity 0.88)
- Automatic Discovery, Association Estimation and Learning of Semantic Attributes for a Thousand Categories (CVPR 2017 · similarity 0.88)
- Joint Unsupervised Learning of Deep Representations and Image Clusters (CVPR 2016 · similarity 0.88)
- Unveiling the Unseen: Identifiable Clusters in Trained Depthwise Convolutional Kernels (ICLR 2024 · 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.