Research map / Vision-language and multimodal / Human and embodied modeling
Nonparametric Deep Fine-grained Clustering with Low-Rank Guided Vision-Language Model
Xulun Ye, Benyu Wu, Jie Hong, Kun Zhou
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 low-rank guidance loss from teacher top-k targets (ablation) | DPMM clustering baseline · Oxford Flower | ACC: 24.4 -> 39.6 | Improved |
| add low-rank guidance loss from teacher top-k targets (ablation) | DPMM clustering baseline · Oxford Flower | NMI: 52.1 -> 75.8 | Improved |
| add input noise perturbation to guidance variant (ablation) | Baseline + Lguidance · Oxford Flower | ACC: 39.6 -> 42.1 | Improved |
| add orthogonal guidance on prototypes (ablation) | DPMM clustering baseline · Oxford Flower | ACC: 24.4 -> 57.5 | Improved |
| combine low-rank guidance, perturbation and orthogonal guidance (ablation) | DPMM clustering baseline · Oxford Flower | ACC: 57.5 -> 65.5 | Improved |
| use CLIP ViT-B/16 teacher instead of ResNet-50 teacher (ablation) | Ours · CUB-200-2011 | ACC: 33.2 -> 41.8 | Improved |
| use CLIP ViT-B/16 teacher instead of ResNet-50 teacher (ablation) | Ours · Oxford Flower | ACC: 65.5 -> 72.6 | Improved |
| increase top-k from 3 to 5 (ablation) | Ours · Stanford Dogs | ACC: 40.3 -> 37.4 | Hurt |
| increase top-k from 3 to 5 (ablation) | Ours · Oxford Flower | ACC: 65.5 -> 66.4 | No clear effect |
| use Ours+CLIP instead of CLUDI | ImageNet200 · ImageNet200 | ACC: 73.7 -> 72.3 | Hurt |
| use Ours+CLIP instead of CLUDI | ImageNet50 · ImageNet50 | ACC: 82.1 -> 84.2 | Improved |
Closest papers by meaning
- Deep Semantic Clustering by Partition Confidence Maximisation (CVPR 2020 · similarity 0.88)
- Language-Assisted Image Clustering Guided by Discriminative Relational Signals and Adaptive Semantic Centers (NeurIPS 2026 · similarity 0.88)
- Fine-R1: Make Multi-modal LLMs Excel in Fine-Grained Visual Recognition by Chain-of-Thought Reasoning (ICLR 2026 · similarity 0.87)
- Democratizing Fine-grained Visual Recognition with Large Language Models (ICLR 2024 · similarity 0.87)
- Image Clustering via the Principle of Rate Reduction in the Age of Pretrained Models (ICLR 2024 · similarity 0.87)
- Benchmarking Large Vision-Language Models on Fine-Grained Image Tasks: A Comprehensive Evaluation (ICLR 2026 · 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.
