Research map / Label-efficient and robust learning / Recognition and long-tail
Online Collaborative Learning for Open-Vocabulary Visual Classifiers
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 |
|---|---|---|---|
| use collaborative label and image embeddings with online updates instead of CNN classification | visual classifiers trained on SBU · NUSWIDE test | mAP%: 24.3 -> 23.9 | Hurt |
| use collaborative label and image embeddings with online updates instead of IncSVM | visual classifiers trained on SBU · CCV test | mAP%: 39.1 -> 40.5 | Helped |
| use collaborative label and image embeddings with online updates instead of LEML | visual classifiers trained on SBU · Flickr30K | mAP%: 2.28 -> 2.48 | Helped |
| use collaborative label and image embeddings with online updates instead of LEML | visual classifiers trained on SBU · COCO validation | mAP%: 3.40 -> 3.52 | Helped |
| use online OCL updates rather than batch optimization (ablation) | OCL · NUSWIDE test | mAP%: 22.7 -> 23.9 | Helped |
| use online OCL updates rather than batch optimization (ablation) | OCL · CCV test | mAP%: 33.9 -> 40.5 | Helped |
| use online OCL updates rather than batch optimization (ablation) | OCL · Flickr30K | mAP%: 1.97 -> 2.48 | Helped |
| use online OCL updates rather than batch optimization (ablation) | OCL · COCO validation | mAP%: 3.11 -> 3.52 | Helped |
| use collaborative latent representations instead of independent N-way classifiers | visual classifiers trained from weakly labeled Web images · NUSWIDE, CCV, Flickr30K, and COCO | mAP% | Mixed |
| use online updates instead of multiple training passes | CNN and IncSVM · four benchmarks | mAP% | Hurt |
| increase the embedding dimension across the tested values (ablation) | OCL, WSABIE, DeViSE, and LEML · four benchmarks | mAP% | Mixed |
| use minibatch OCL updates rather than single-sample updates (ablation) | OCL · four benchmarks | mAP% | Helped |
Closest papers by meaning
- Learning to Learn Image Classifiers With Visual Analogy (CVPR 2019 · similarity 0.87)
- OpenMix: Reviving Known Knowledge for Discovering Novel Visual Categories in an Open World (CVPR 2021 · similarity 0.87)
- Pre-Trained Vision-Language Models as Noisy Partial Annotators (AAAI 2025 · similarity 0.85)
- OVMR: Open-Vocabulary Recognition with Multi-Modal References (CVPR 2024 · similarity 0.85)
- Modeling Collaborator: Enabling Subjective Vision Classification With Minimal Human Effort via LLM Tool-Use (CVPR 2024 · similarity 0.85)
- Correlative and Discriminative Label Grouping for Multi-Label Visual Prompt Tuning (CVPR 2025 · 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.