Research map / Vision-language and multimodal / CLIP and zero-shot
ARGENT: Adaptive Hierarchical Image-Text Representations
Chuong Huynh, Hossein Souri, Abhinav Kumar, Vitali Petsiuk, Deen Dayal Mohan, Suren Kumar
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 |
|---|---|---|---|
| replace vanilla entailment loss with AdaEnt (ablation) | MERU-S · CC3M ablation | PEP AUC: 73.5 -> 95.5 | Improved |
| replace vanilla entailment loss with AdaEnt (ablation) | MERU-S · CC3M ablation | PEP AP: 45.0 -> 60.9 | Improved |
| replace vanilla entailment loss with AdaEnt (ablation) | HyCoCLIP-S · CC3M ablation | PEP AUC: 94.6 -> 97.3 | Improved |
| replace vanilla entailment loss with AdaEnt (ablation) | HyCoCLIP-B · CC3M ablation | PEP AUC: 94.2 -> 97.9 | Improved |
| replace vanilla entailment loss with AdaEnt (ablation) | HyCoCLIP-L · CC3M ablation | PEP AP: 76.0 -> 79.2 | Improved |
| replace vanilla entailment loss with AdaEnt (ablation) | HyCoCLIP-S · CC3M ablation | classification accuracy: 22.9 -> 22.8 | No clear effect |
| apply adaptive weight h only to intra-modality pairs rather than all pairs (ablation) | HyCoCLIP-S · CC3M ablation | PEP AUC: 71.1 -> 97.3 | Improved |
| apply adaptive weight h to intra-modality pairs vs none (ablation) | HyCoCLIP-S · CC3M ablation | PEP AUC: 94.9 -> 97.3 | Improved |
| add regularization loss L_Reg (ablation) | HyCoCLIP-S · CC3M ablation | PEP AUC: 96.2 -> 97.3 | Improved |
| add regularization loss L_Reg (ablation) | MERU-S · CC3M ablation | classification accuracy: 20.8 -> 20.6 | Mixed |
| replace HyCoCLIP entailment loss with AdaEnt (ARGENT) | HyCoCLIP-S · ImageNet zero-shot | accuracy: 37.3 -> 38.8 | Improved |
| replace HyCoCLIP entailment loss with AdaEnt (ARGENT) | HyCoCLIP-S · 16-dataset zero-shot avg | accuracy: 39.2 -> 38.4 | Hurt |
Closest papers by meaning
- Compositional Entailment Learning for Hyperbolic Vision-Language Models (ICLR 2025 · similarity 0.93)
- Understanding Fine-tuning CLIP for Open-vocabulary Semantic Segmentation in Hyperbolic Space (CVPR 2025 · similarity 0.89)
- HyperET: Efficient Training in Hyperbolic Space for Multi-modal Large Language Models (NeurIPS 2025 · similarity 0.88)
- Not Only Text: Exploring Compositionality of Visual Representations in Vision-Language Models (CVPR 2025 · similarity 0.87)
- Exploiting the Asymmetric Uncertainty Structure of Pre-trained VLMs on the Unit Hypersphere (NeurIPS 2025 · similarity 0.87)
- DH-Set: Improving Vision-Language Alignment with Diverse and Hybrid Set-Embeddings Learning (CVPR 2025 · similarity 0.87)
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.
