DPGN: Distribution Propagation Graph Network for Few-Shot Learning

Lin, Yang; Li, Liangliang; Zhang, Zilun; Zhou, Xinyu; Zhou, Erjin; Liu, Yu

doi:10.1109/cvpr42600.2020.01340

Cited by 212 publications

(129 citation statements)

References 12 publications

Supporting

Mentioning

129

Contrasting

Order By: Relevance

“…Transductive few-shot methods [1], [5], [12], [26], [31], [50], [59] assume that the model simultaneously accesses all the query set. A transductive episodic-wise adaptive metric (TEAM) [31] defined the optimization process as a standard semi-definite programming problem to train a generalizable classifier.…”

Section: Related Workmentioning

confidence: 99%

“…A transductive episodic-wise adaptive metric (TEAM) [31] defined the optimization process as a standard semi-definite programming problem to train a generalizable classifier. A distribution propagation graph network (DPGN) [50] proposed utilizing both the distribution-level and instance-level relations by designing a dual complete graph network consisting of a point graph and a distribution graph. [5] proposed transductive fine-tuning, which pursues outputs with a peaked posterior or low Shannon entropy, and a hardness metric to deliver a standardized evaluation protocol.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Augmenting Few-Shot Learning With Supervised Contrastive Learning

Lee

Yoo

2021

IEEE Access

View full text Add to dashboard Cite

Few-shot learning deals with a small amount of data which incurs insufficient performance with conventional cross-entropy loss. We propose a pretraining approach for few-shot learning scenarios. That is, considering that the feature extractor quality is a critical factor in few-shot learning, we augment the feature extractor using a contrastive learning technique. It is reported that supervised contrastive learning applied to base class training in transductive few-shot training pipeline leads to improved results, outperforming the state-of-the-art methods on Mini-ImageNet and CUB. Furthermore, our experiment shows that a much larger dataset is needed to retain few-shot classification accuracy when domain-shift degradation exists, and if our method is applied, the need for a large dataset is eliminated. The accuracy gain can be translated to a runtime reduction of 3.87× in a resource-constrained environment.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Augmenting Few-Shot Learning With Supervised Contrastive Learning

Lee

Yoo

2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…A learnable scaling parameter is added to the metric function, which greatly improves the classification performance. GNN [38] EGNN [39]DPGN [40] These articles introduced graph network to transfer sample distance measurement from European space to non-European space to achieve target classification.…”

Section: Prototypementioning

confidence: 99%

Feature Transformation Network for Few-Shot Learning

Wang

Zhou

2021

IEEE Access

View full text Add to dashboard Cite

Few-shot learning researches to learn a novel concept from a handful of labeled samples. Due to the small amount of training data, deep network has the risk of over-fitting. Although many previous approaches based on metric criterion can make significant progress to tackle this challenge, they not only ignore the association between query set and support set when learning sample representation, but also fail to focus greater attention in the target area. To cope with these issues, we propose a novel feature transformation network (FTN) for few-shot image classification. Specifically, to draw inferences about other instances from only a few examples, it is expected to learn a model that has more discriminative representation of the target attributes and robust generalization ability. To this end, we introduce an attention-based affinity matrix to transform the semantical enhanced embedding vectors of query samples by associating the support set, thereby guiding the network to learn a sample representation that embodies higher semantic information in the target area. Furthermore, aiming at highlighting the object region in the feature maps, and strengthening the pertinence of similarity measurement between samples, a global and local feature fusion module is designed to fuse the support set samples features. The comprehensive experiments validate the doable of our model, and our method achieves the state-of-the-art performance on two public benchmark datasets, namely, general object dataset mini-ImageNet and fine-grained dataset Caltech-UCSD Birds-200-2011 (CUB). INDEX TERMS Few-shot learning, feature transformation, feature fusion, metric criterion.

show abstract

“…Metric learning based methods [11,5,12,13] finds better similarity metrics for image embeddings. Graph based methods [14,9] construct a complete graph that connects all samples of a fewshot task to perform information propagation and the queries are classified based on their relations to shots. Besides, there are also image retrieval based works [15] and reinforcement learning based ones [16].…”

Section: Related Workmentioning

confidence: 99%

“…The performance is reported as the mean accuracy and the 95% confidence interval. Usually, the training set is also sampled into many training few-shot tasks to optimize the model [5,8,9]. Recently, subspace learning based methods for few-shot learning [6,7] draws increasing attention.…”

Section: Introductionmentioning

confidence: 99%

GSS: Graph-Based Subspace Learning with Shots Initialization for few-shot Recognition

Wang

Zhang

Liu

2021

2021 IEEE International Conference on Multimedia and Expo (ICME)

View full text Add to dashboard Cite

Previous methods of few-shot Learning mostly solve different few-shot recognition tasks in an identical feature space. But identical features are hard to fit various tasks. Some works show that learning a unique subspace for each few-shot recognition task can improve the signal-noise ratio (SNR) of the features and boost the performance. However, there are still two problems remaining. First, in constructing the subspace for few-shot task, often some information (embeddings of queries or labels of shots) are discarded. Second, the eigendecomposition of covariance matrix is usually needed, which degrades the efficiency of the whole model. In this paper, we propose Graph-based Subspace learning with Shots initialization (GSS) for few-shot recognition to learn a better subspace efficiently. In GSS, the bases of the subspace are directly initialized with labels based on shots (given labeled samples) and iteratively updated for better discrimination based on a graph that connects bases and all samples. Extensive experiments on four few-shot benchmark datasets show that GSS reports better performance and higher efficient compared with previous subspace based methods and achieves state-of-theart performance.

show abstract

DPGN: Distribution Propagation Graph Network for Few-Shot Learning

Cited by 212 publications

References 12 publications

Augmenting Few-Shot Learning With Supervised Contrastive Learning

Augmenting Few-Shot Learning With Supervised Contrastive Learning

Feature Transformation Network for Few-Shot Learning

GSS: Graph-Based Subspace Learning with Shots Initialization for few-shot Recognition

Contact Info

Product

Resources

About