HyperAdapter performs PEFT of ViTs via soft hypergraph construction, hyperedge-level bottleneck adaptation, and incidence-based diffusion, claiming consistent gains over token-wise adapters on structured visual benchmarks.
arXiv preprint arXiv:2408.11351 (2024)
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
A new semi-supervised hypergraph Concept Bottleneck Model framework improves label efficiency and interpretability for medical image diagnosis on PAS ultrasound, breast ultrasound, and SkinCon datasets.
citing papers explorer
-
Structured Hyperedge Adaptation for Parameter-Efficient Fine-Tuning of Vision Transformers
HyperAdapter performs PEFT of ViTs via soft hypergraph construction, hyperedge-level bottleneck adaptation, and incidence-based diffusion, claiming consistent gains over token-wise adapters on structured visual benchmarks.
-
Learning Label-Efficient Interpretable Medical Image Diagnosis via Semi-supervised Hypergraph Concept Bottleneck Model
A new semi-supervised hypergraph Concept Bottleneck Model framework improves label efficiency and interpretability for medical image diagnosis on PAS ultrasound, breast ultrasound, and SkinCon datasets.