Research map / Vision-language and multimodal / Cross-modal learning
MAGIC: Multi-Granularity Language-Informed Image Clustering
Xiaohan Zhang, Chao Zhang, Chunlin Chen, Huaxiong Li
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 multi-modal language information (MAGIC vs k-means on image CLIP) (ablation) | MAGIC · Five-dataset average | AVG: 65.6 -> 87.2 | Improved |
| use VLM descriptions instead of WordNet nouns (ablation) | MAGIC · Five-dataset average | AVG: 84.3 -> 87.2 | Improved |
| add balance loss L_BL (ablation) | MAGIC · CIFAR-20 | ACC: 27.8 -> 66.1 | Improved |
| add cross-entropy alignment loss L_CE (ablation) | MAGIC · ImageNet-Dogs | ACC: 90.1 -> 93.1 | Improved |
| add inter-modal contrastive distillation to intra-modal (ablation) | MAGIC · CIFAR-20 | ACC: 55.6 -> 66.1 | Improved |
| fuse fine and coarse text embeddings instead of fine only (ablation) | MAGIC · CIFAR-20 | ACC: 61.0 -> 66.1 | Improved |
| add semantic adapters (image and text) (ablation) | MAGIC · CIFAR-20 | ACC: 57.1 -> 66.1 | Improved |
| add residual connection to semantic adapters (ablation) | MAGIC · CIFAR-20 | ACC: 66.1 -> 59.9 | Hurt |
| bottleneck ratio 1/4 vs 1 or 1/8 (ablation) | MAGIC · CIFAR-20 | ACC: 62.9 -> 66.1 | Improved |
| use cross-granularity attention instead of addition or concatenation (ablation) | MAGIC · ImageNet-Dogs, CIFAR-20 | clustering metrics | Improved |
| concatenate image embedding into text cluster head (consensus representation) (ablation) | MAGIC · ImageNet-Dogs, CIFAR-20 | clustering metrics | Improved |
| use Qwen2.5-VL instead of BLIP-2 or WordNet as language generator (ablation) | MAGIC · ImageNet-Dogs, CIFAR-20 | clustering metrics | Improved |
Closest papers by meaning
- Language-Assisted Image Clustering Guided by Discriminative Relational Signals and Adaptive Semantic Centers (NeurIPS 2026 · similarity 0.91)
- MaVEn: An Effective Multi-granularity Hybrid Visual Encoding Framework for Multimodal Large Language Model (NeurIPS 2024 · similarity 0.86)
- Benchmarking Large Vision-Language Models on Fine-Grained Image Tasks: A Comprehensive Evaluation (ICLR 2026 · similarity 0.86)
- Multi-Modal Proxy Learning Towards Personalized Visual Multiple Clustering (CVPR 2024 · similarity 0.86)
- Towards Semantic Equivalence of Tokenization in Multimodal LLM (ICLR 2025 · similarity 0.85)
- Granulon: Awakening Pixel-Level Visual Encoders with Adaptive Multi-Granularity Semantics for MLLM (CVPR 2026 · 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.
