Adaptive Subspaces for Few-Shot Learning

Simon, Christian; Koniusz, Piotr; Nock, Richard; Harandi, Mehrtash

doi:10.1109/cvpr42600.2020.00419

Cited by 367 publications

(206 citation statements)

References 24 publications

Supporting

Mentioning

197

Contrasting

Order By: Relevance

“…ProtoNet uses Euclidean distance while RelationNet compares an embedding f ϕ and query samples using an additional parameterized CNN-based 'relation module'. MetaOptNet [48] and DSN-MR [49] are also metric-based approaches. MetaOptNet provides an end-to-end method with regularized linear classifiers i.e., ridge regression and SVM.…”

Section: Results and Comparisonsmentioning

confidence: 99%

“…The linear classifiers can learn better class boundaries using negative examples at a modest increase in computational costs. Simon et al [49] observed that high-order information is preferred over low-order to improve the classifier's capability in the low data regime; hence one hopes a subspace method can form a robust classifier. The authors develop a dynamic classifier that computes a subspace of feature space for each category, and the features of query samples are projected into the subspace for comparison.…”

Section: Few-shot Classification Via Meta-learningmentioning

confidence: 99%

“…All our code was implemented in Pytorch and run with 4 NVIDIA RTX 2080 Ti. Note that, training in DSN-MR [49] with a ResNet-12 backbone requires 4 GPUs with ∼10 GB/GPU.…”

Section: Implementation Detailsmentioning

confidence: 99%

“…MathingNets [45] 64-64-64-64 2016 cosine similarity ProtoNet [32] 64-64-64-64 2017 Euclidean distance RelationNet [35] 64-96-128-256 2018 additional CNN relation module TADAM [46] ResNet-12 2018 scaled Euclidean distance, task-specific MetaOptNet [48] ResNet-12 2019 ridge regression, SVM DSN-MR [49] ResNet-12 2020 subspace, SVD…”

Section: Metric-basedmentioning

confidence: 99%

See 3 more Smart Citations

Few-Shot Classification of Aerial Scene Images via Meta-Learning

et al. 2020

View full text Add to dashboard Cite

Convolutional neural network (CNN) based methods have dominated the field of aerial scene classification for the past few years. While achieving remarkable success, CNN-based methods suffer from excessive parameters and notoriously rely on large amounts of training data. In this work, we introduce few-shot learning to the aerial scene classification problem. Few-shot learning aims to learn a model on base-set that can quickly adapt to unseen categories in novel-set, using only a few labeled samples. To this end, we proposed a meta-learning method for few-shot classification of aerial scene images. First, we train a feature extractor on all base categories to learn a representation of inputs. Then in the meta-training stage, the classifier is optimized in the metric space by cosine distance with a learnable scale parameter. At last, in the meta-testing stage, the query sample in the unseen category is predicted by the adapted classifier given a few support samples. We conduct extensive experiments on two challenging datasets: NWPU-RESISC45 and RSD46-WHU. The experimental results show that our method yields state-of-the-art performance. Furthermore, several ablation experiments are conducted to investigate the effects of dataset scale, the impact of different metrics and the number of support shots; the experiment results confirm that our model is specifically effective in few-shot settings.

show abstract

Section: Results and Comparisonsmentioning

confidence: 99%

Section: Few-shot Classification Via Meta-learningmentioning

confidence: 99%

“…All our code was implemented in Pytorch and run with 4 NVIDIA RTX 2080 Ti. Note that, training in DSN-MR [49] with a ResNet-12 backbone requires 4 GPUs with ∼10 GB/GPU.…”

Section: Implementation Detailsmentioning

confidence: 99%

Section: Metric-basedmentioning

confidence: 99%

See 2 more Smart Citations

Few-Shot Classification of Aerial Scene Images via Meta-Learning

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Most of the early FSL methods [ 38 , 40 , 42 , 45 ] utilized a four-layer convolutional network (Conv-4) as the embedding backbone, while more recent models found that such a shallow embedding network might lead to underfitting. In this work, we take ResNet-12, the most popular backbone in current FSL literature [ 51 , 52 , 59 ], as our embedding network. As illustrated in Figure 4 , the ResNet-12 is a smaller version of the ResNet [ 9 ], containing four residual blocks and generates 512-dimensional embeddings after a global average pooling (GAP).…”

Section: Methodsmentioning

confidence: 99%

RS-SSKD: Self-Supervision Equipped with Knowledge Distillation for Few-Shot Remote Sensing Scene Classification

Zhang

Wang

et al. 2021

Sensors

View full text Add to dashboard Cite

While growing instruments generate more and more airborne or satellite images, the bottleneck in remote sensing (RS) scene classification has shifted from data limits toward a lack of ground truth samples. There are still many challenges when we are facing unknown environments, especially those with insufficient training data. Few-shot classification offers a different picture under the umbrella of meta-learning: digging rich knowledge from a few data are possible. In this work, we propose a method named RS-SSKD for few-shot RS scene classification from a perspective of generating powerful representation for the downstream meta-learner. Firstly, we propose a novel two-branch network that takes three pairs of original-transformed images as inputs and incorporates Class Activation Maps (CAMs) to drive the network mining, the most relevant category-specific region. This strategy ensures that the network generates discriminative embeddings. Secondly, we set a round of self-knowledge distillation to prevent overfitting and boost the performance. Our experiments show that the proposed method surpasses current state-of-the-art approaches on two challenging RS scene datasets: NWPU-RESISC45 and RSD46-WHU. Finally, we conduct various ablation experiments to investigate the effect of each component of the proposed method and analyze the training time of state-of-the-art methods and ours.

show abstract

Fighting fire with fire: A spatial–frequency ensemble relation network with generative adversarial learning for adversarial image classification

et al. 2021

View full text Add to dashboard Cite

Adversarial images generated by generative adversarial networks are not close to any existing benign images, and contain nonrobust features that have been identified as critical to the robustness of a machine learning model. Since adversarial images have an underlying distribution that differs from normal images, these kinds of images can offer valuable features for training a robust model. To deal with these special features, we focus on a novel machine learning task of adversarial images classification, where adversarial images can be used to investigate the problem of classifying adversarial images themselves. In the setting of this novel task, adversarial images are the ONLY kind of data used in training and testing, rather than not just a set of testing images as usual. To this end, we propose a novel spatial–frequency ensemble relation network with generative adversarial learning. First, we present a spatial–frequency ensemble representation learning to extract the feature of training images. Second, we design a meta‐learning‐based relation model to gain the relationship between images. Third, to achieve a robust model, we utilize generative adversarial learning and transform the relationship into a Jacobian matrix. Finally, we design a discriminator model that determines whether an adversarial image is from the matching category or not. Experimental results demonstrate that our approach achieves significantly higher performance compared with other state‐of‐the‐arts.

show abstract

Adaptive Subspaces for Few-Shot Learning

Cited by 367 publications

References 24 publications

Few-Shot Classification of Aerial Scene Images via Meta-Learning

Few-Shot Classification of Aerial Scene Images via Meta-Learning

RS-SSKD: Self-Supervision Equipped with Knowledge Distillation for Few-Shot Remote Sensing Scene Classification

Fighting fire with fire: A spatial–frequency ensemble relation network with generative adversarial learning for adversarial image classification

Contact Info

Product

Resources

About