FDSML:公式驅動監督式度量學習與參數感知三元組損失
FDSML: Formula-Driven Supervised Metric Learning with Parameter-Aware Triplet Loss
FDSML提出以生成參數提供關係監督的度量學習方法,結合 FineFractalDB 和 PA-Triplet Loss,利用迭代函數系統參數空間中的距離設定正、負樣本,訓練影像嵌入向量。
Published on Oct 29
Authors:
,
,
,
,
Abstract
Formula-driven supervised metric learning uses continuous generative parameters to train embeddings with parameter-aware triplet loss, improving synthetic pre-training for image retrieval and robustness.
Formula-Driven Supervised Learning (FDSL) enables vision pre-training without real images, but existing methods cast pre-training as classification on discretized labels, discarding continuity in the generative process. We propose Formula-Driven Supervised Metric Learning (FDSML), which derives relational supervision from image-generation parameters beyond discrete class labels. FDSML combines FineFractalDB, our refined fractal dataset with higher intra-class variation and more stable structures, with a Parameter-Aware Triplet (PA-Triplet) Loss that assigns positives and negatives according to distances in the underlying Iterated Function System parameter space. This objective trains embeddings to respect parameter-derived relative similarity orderings. Experiments show that FDSML improves transferability to image retrieval while remaining competitive on downstream classification and improving robustness under several corruption settings. On ResNet-50, Recall@1 improves from 65.6% to 74.0% on Stanford Online Products and from 56.2% to 78.4% on Stanford Cars over FDSL pre-training. These results suggest that parameter-derived relational supervision provides a stronger synthetic pre-training signal than label-based classification for transfer beyond recognition.
View arXiv page View PDF Project page GitHub 0 Add to collection
Community
Paper author
Paper submitter
FDSML pre-trains vision models without real images, using the generating formula itself as supervision: fractals with more similar parameters should lie closer in the embedding space. We turn this parameter-space ordering into triplets, so the model learns how images relate to each other.
Get this paper in your agent:
hf papers read 2610.39423
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash
Models citing this paper 12
Image Feature Extraction • Updated about 8 hours agoFDSML/fdsml-finefractaldb-resnet50-all
Image Feature Extraction • Updated about 8 hours agoFDSML/fdsml-finefractaldb-resnet50-diff
Image Feature Extraction • Updated about 8 hours agoFDSML/fdsml-finefractaldb-resnet50-each
Image Feature Extraction • Updated about 8 hours agoFDSML/fdsml-finefractaldb-resnet50-same
Datasets citing this paper 1
FDSML/FineFractalDB
Viewer • Updated about 8 hours ago • 1M • 68
Spaces citing this paper 0
No Space linking this paper
Cite arxiv.org/abs/2610.39423 in a Space README.md to link it from this page.
Collections including this paper 0
No Collection including this paper
Add this paper to a collection to link it from this page.
來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co