Few-shot Learning Approach Research Articles

Few-shot deep learning for AFM force curve characterization of single-molecule interactions

Deep learning (DL)-based analytics has the scope to transform the field of atomic force microscopy (AFM) with regard to fast and bias-free measurement characterization. For example, AFM force-distance curves can help estimate important parameters of binding kinetics, such as the most probable rupture force, binding probability, association, and dissociation constants, as well as receptor density on live cells. Other than the ideal single-rupture event in the force-distance curves, there can be no-rupture, double-rupture, or multiple-rupture events. The current practice is to go through such datasets manually, which can be extremely tedious work for the experimentalists. We address this issue by adopting a few-shot learning approach to build sample-efficient DL models that demonstrate better performance than shallow ML models while matching the performance of moderately trained humans. We also release our AFM force curve dataset and annotations publicly as a benchmark for the research community.

SAPENet: Self-Attention based Prototype Enhancement Network for Few-shot Learning

Few-shot learning considers the problem of learning unseen categories given only a few labeled samples. As one of the most popular few-shot learning approaches, Prototypical Networks have received considerable attention owing to their simplicity and efficiency. However, a class prototype is typically obtained by averaging a few labeled samples belonging to the same class, which treats the samples as equally important and is thus prone to learning redundant features. Herein, we propose a self-attention based prototype enhancement network (SAPENet) to obtain a more representative prototype for each class. SAPENet utilizes multi-head self-attention mechanisms to selectively augment discriminative features in each sample feature map, and generates channel attention maps between intra-class sample features to attentively retain informative channel features for that class. The augmented feature maps and attention maps are finally fused to obtain representative class prototypes. Thereafter, a local descriptor-based metric module is employed to fully exploit the channel information of the prototypes by searching k similar local descriptors of the prototype for each local descriptor in the unlabeled samples for classification. We performed experiments on multiple benchmark datasets: miniImageNet, tieredImageNet, and CUB-200-2011. The experimental results on these datasets show that SAPENet achieves a considerable improvement compared to Prototypical Networks and also outperforms related state-of-the-art methods.

Introducing Urdu Digits Dataset with Demonstration of an Efficient and Robust Noisy Decoder-Based Pseudo Example Generator

In the present work, we propose a novel method utilizing only a decoder for generation of pseudo-examples, which has shown great success in image classification tasks. The proposed method is particularly constructive when the data are in a limited quantity used for semi-supervised learning (SSL) or few-shot learning (FSL). While most of the previous works have used an autoencoder to improve the classification performance for SSL, using a single autoencoder may generate confusing pseudo-examples that could degrade the classifier’s performance. On the other hand, various models that utilize encoder–decoder architecture for sample generation can significantly increase computational overhead. To address the issues mentioned above, we propose an efficient means of generating pseudo-examples by using only the generator (decoder) network separately for each class that has shown to be effective for both SSL and FSL. In our approach, the decoder is trained for each class sample using random noise, and multiple samples are generated using the trained decoder. Our generator-based approach outperforms previous state-of-the-art SSL and FSL approaches. In addition, we released the Urdu digits dataset consisting of 10,000 images, including 8000 training and 2000 test images collected through three different methods for purposes of diversity. Furthermore, we explored the effectiveness of our proposed method on the Urdu digits dataset by using both SSL and FSL, which demonstrated improvement of 3.04% and 1.50% in terms of average accuracy, respectively, illustrating the superiority of the proposed method compared to the current state-of-the-art models.

Few-shot learning approach with multi-scale feature fusion and attention for plant disease recognition.

Image-based deep learning method for plant disease diagnosing is promising but relies on large-scale dataset. Currently, the shortage of data has become an obstacle to leverage deep learning methods. Few-shot learning can generalize to new categories with the supports of few samples, which is very helpful for those plant disease categories where only few samples are available. However, two challenging problems are existing in few-shot learning: (1) the feature extracted from few shots is very limited; (2) generalizing to new categories, especially to another domain is very tough. In response to the two issues, we propose a network based on the Meta-Baseline few-shot learning method, and combine cascaded multi-scale features and channel attention. The network takes advantage of multi-scale features to rich the feature representation, uses channel attention as a compensation module efficiently to learn more from the significant channels of the fused features. Meanwhile, we propose a group of training strategies from data configuration perspective to match various generalization requirements. Through extensive experiments, it is verified that the combination of multi-scale feature fusion and channel attention can alleviate the problem of limited features caused by few shots. To imitate different generalization scenarios, we set different data settings and suggest the optimal training strategies for intra-domain case and cross-domain case, respectively. The effects of important factors in few-shot learning paradigm are analyzed. With the optimal configuration, the accuracy of 1-shot task and 5-shot task achieve at 61.24% and 77.43% respectively in the task targeting to single-plant, and achieve at 82.52% and 92.83% in the task targeting to multi-plants. Our results outperform the existing related works. It demonstrates that the few-shot learning is a feasible potential solution for plant disease recognition in the future application.

A novel few-shot learning approach for wind power prediction applying secondary evolutionary generative adversarial network

The accuracy and stability of wind power forecasting are very important for the operation of wind farms. However, for the newly built wind farms without sufficient historical data, it is difficult to make a more accurate prediction. Therefore, it is of great significance to explore a method to improve the wind power prediction accuracy with no sufficient historical data available. In this paper, a novel prediction model is proposed to address the few-shot learning problem of wind power prediction in new-built wind farms based on secondary evolutionary generative adversarial networks (SEGAN) and dual-dimension attention mechanism (DDAM) assisted bidirectional gate recurrent unit (BiGRU). The SEGAN first introduces the secondary evolutionary learning paradigm into learning GAN, aiming to learn the marginal distribution of real data and generate high-quality realistic data to augment the training dataset. In the prediction stage, the DDAM is attempted to obtain a new input matrix with global weight allocation and improve the sensitivity of the BiGRU model to the key information of the input data. The proposed SEGAN-DDAM-BiGRU model is validated on the data from the Galicia Wind Farm in Sotavento and the experimental results show that the proposed model is applicative for short-term prediction of new-built wind farms.

A few-shot malware classification approach for unknown family recognition using malware feature visualization

With the ever-increasing threat of malware attacks, building an effective malware classifier to detect malware promptly is of utmost importance. Malware visualization approaches and deep learning techniques have proven effective in classifying sophisticated malware from benchmark datasets. A major problem with traditional deep learning classifier is the need to re-train the classifier when a new malware family emerges. In this paper, we propose few-shot classification techniques which allows us to classify malware based on a few instances and without the need for re-training the classifier for novel malware families. We also propose a novel malware visualization technique that can represent a malware binary as a 3-channel image. We experiment with two distinct few-shot learning architectures namely CSNN (Convolutional Siamese Neural Network) and Shallow-FS (Shallow Few-Shot). CSNN is more suitable when scarce data is available for training, otherwise Shallow-FS can be used to achieve better performance. Our architectures outperforms state of the art few-shot learning approaches and achieves high accuracy in traditional malware classification. Our experiments show our models’ ability to classify recent and novel malware families from just a few instances with high accuracy.

Insect Pest Image Recognition: A Few-Shot Machine Learning Approach including Maturity Stages Classification

Recognizing insect pests using images is an important and challenging research issue. A correct species classification will help choosing a more proper mitigation strategy regarding crop management, but designing an automated solution is also difficult due to the high similarity between species at similar maturity stages. This research proposes a solution to this problem using a few-shot learning approach. First, a novel insect data set based on curated images from IP102 is presented. The IP-FSL data set is composed of 97 classes of adult insect images, and 45 classes of early stages, totalling 6817 images. Second, a few-shot prototypical network is proposed based on a comparison with other state-of-art models and further divergence analysis. Experiments were conducted separating the adult classes and the early stages into different groups. The best results achieved an accuracy of 86.33% for the adults, and 87.91% for early stages, both using a Kullback–Leibler divergence measure. These results are promising regarding a crop scenario where the more significant pests are few and it is important to detect them at earlier stages. Further research directions would be in evaluating a similar approach in particular crop ecosystems, and testing cross-domains.

Towards a Deep-Learning Approach for Prediction of Fractional Flow Reserve from Optical Coherence Tomography

Cardiovascular disease (CVD) is the number one cause of death worldwide, and coronary artery disease (CAD) is the most prevalent CVD, accounting for 42% of these deaths. In view of the limitations of the anatomical evaluation of CAD, Fractional Flow Reserve (FFR) has been introduced as a functional diagnostic index. Herein, we evaluate the feasibility of using deep neural networks (DNN) in an ensemble approach to predict the invasively measured FFR from raw anatomical information that is extracted from optical coherence tomography (OCT). We evaluate the performance of various DNN architectures under different formulations: regression, classification—standard, and few-shot learning (FSL) on a dataset containing 102 intermediate lesions from 80 patients. The FSL approach that is based on a convolutional neural network leads to slightly better results compared to the standard classification: the per-lesion accuracy, sensitivity, and specificity were 77.5%, 72.9%, and 81.5%, respectively. However, since the 95% confidence intervals overlap, the differences are statistically not significant. The main findings of this study can be summarized as follows: (1) Deep-learning (DL)-based FFR prediction from reduced-order raw anatomical data is feasible in intermediate coronary artery lesions; (2) DL-based FFR prediction provides superior diagnostic performance compared to baseline approaches that are based on minimal lumen diameter and percentage diameter stenosis; and (3) the FFR prediction performance increases quasi-linearly with the dataset size, indicating that a larger train dataset will likely lead to superior diagnostic performance.

Cross-Domain Few-Shot Learning Approach for Lithium-Ion Battery Surface Defects Classification Using an Improved Siamese Network

It is difficult to detect the surface defects of a lithium battery with an aluminum/steel shell. The reflectivity, lack of 3D information on the battery surface, and the shortage of many datasets make the 2D detection method hard to apply in this field. In this paper, a cross-domain few-shot learning (FSL) approach for lithium-ion battery defect classification using an improved siamese network (BSR-SNet) is proposed. To obtain the critical 3D surface of the lithium-ion battery, a multiexposure-based structured light method is utilized. Then, the heights of the 3D cloud points are transferred to grayscale information and are saved as 8-bit 2D images. For the FSL task, the DAGM 2007 datasets are used as the source domain to pre-train the improved siamese model. To avoid negative mitigation in the target domain, batch spectral regularization (BSR) is added as a penalizer in the loss function. The accuracies of the experimental results are 93.3% for 10-shot batteries and 91.0% for 5-shot batteries, which means that our method can be used to classify the surface defects of lithium batteries well.

OptNCMiner: a deep learning approach for the discovery of natural compounds modulating disease-specific multi-targets

BackgroundDue to their diverse bioactivity, natural product (NP)s have been developed as commercial products in the pharmaceutical, food and cosmetic sectors as natural compound (NC)s and in the form of extracts. Following administration, NCs typically interact with multiple target proteins to elicit their effects. Various machine learning models have been developed to predict multi-target modulating NCs with desired physiological effects. However, due to deficiencies with existing chemical-protein interaction datasets, which are mostly single-labeled and limited, the existing models struggle to predict new chemical-protein interactions. New techniques are needed to overcome these limitations.ResultsWe propose a novel NC discovery model called OptNCMiner that offers various advantages. The model is trained via end-to-end learning with a feature extraction step implemented, and it predicts multi-target modulating NCs through multi-label learning. In addition, it offers a few-shot learning approach to predict NC-protein interactions using a small training dataset. OptNCMiner achieved better prediction performance in terms of recall than conventional classification models. It was tested for the prediction of NC-protein interactions using small datasets and for a use case scenario to identify multi-target modulating NCs for type 2 diabetes mellitus complications.ConclusionsOptNCMiner identifies NCs that modulate multiple target proteins, which facilitates the discovery and the understanding of biological activity of novel NCs with desirable health benefits.

A Cluster-then-label Approach for Few-shot Learning with Application to Automatic Image Data Labeling

Few-shot learning (FSL) aims at learning to generalize from only a small number of labeled examples for a given target task. Most current state-of-the-art FSL methods typically have two limitations. First, they usually require access to a source dataset (in a similar domain) with abundant labeled examples, which may not always be possible due to privacy concerns and copyright issues. Second, they typically do not offer any estimation of the generalization error on the target FSL task, because the handful of labeled examples must be used for training and cannot spare a validation subset. In this article, we propose a cluster-then-label approach to perform few-shot learning. Our approach does not require access to the labeled source dataset and provides an estimation of generalization error. We show empirically, on four benchmark datasets, that our approach provides competitive predictive performance to state-of-the-art FSL approaches and our generalization error estimation is accurate. Finally, we explore the application of our proposed method to automatic image data labeling. We compare our method with existing automatic data labeling systems. The end-to-end performance of our method outperforms the state-of-the-art automatic data labeling system Snuba by 26% and is only 7% away from the fully supervised upper bound.

Anomaly detection model based on few-shot learning and memory modules

Anomaly detection is a key issue in public security. Its accuracy is essential to identify abnormalities and take corresponding actions to ensure the safety of relevant objects, which have a broad application space. The traditional anomaly detection method based on deep learning has too strong generalization ability. At the same time, it lacks recognition ability because it only uses normal data for training. To this end, we propose an anomaly detection model based on few-shot learning, guided by memory modules and trained by a large number of normal samples combined with a small number of observed abnormalities. We introduce memory modules to record normal features, which has the function of updating and reading. When training modules, feature compactness, and separateness loss are utilized, it successfully weakens the strong generalization of CNN and improves the memory module to identify normal samples. Then, based on the few-shot learning approach, we learn a more compact normal data distribution and expand the margin between normal and anomalous events to improve the discriminant ability. Many experimental results demonstrate that our method is practical and feasible, and its performance is better than the existing detection methods.

Classification and Fast Few-Shot Learning of Steel Surface Defects with Randomized Network

Quality inspection is inevitable in the steel industry so there are already benchmark datasets for the visual inspection of steel surface defects. In our work, we show, contrary to previous recent articles, that a generic state-of-art deep neural network is capable of almost-perfect classification of defects of two popular benchmark datasets. However, in real-life applications new types of errors can always appear, thus incremental learning, based on very few example shots, is challenging. In our article, we address the problems of the low number of available shots of new classes, the catastrophic forgetting of known information when tuning for new artifacts, and the long training time required for re-training or fine-tuning existing models. In the proposed new architecture we combine EfficientNet deep neural networks with randomized classifiers to aim for an efficient solution for these demanding problems. The classification outperforms all other known approaches, with an accuracy 100% or almost 100%, on the two datasets with the off-the-shelf network. The proposed few-shot learning approach shows considerably higher accuracy at a low number of shots than the different methods under testing, while its speed is significantly (at least 10 times) higher than its competitors. According to these results, the classification and few-shot learning of steel surface defects can be solved more efficiently than was possible before.

An Adaptive Embedding Network with Spatial Constraints for the Use of Few-Shot Learning in Endangered-Animal Detection

Image recording is now ubiquitous in the fields of endangered-animal conservation and GIS. However, endangered animals are rarely seen, and, thus, only a few samples of images of them are available. In particular, the study of endangered-animal detection has a vital spatial component. We propose an adaptive, few-shot learning approach to endangered-animal detection through data augmentation by applying constraints on the mixture of foreground and background images based on species distributions. First, the pre-trained, salient network U2-Net segments the foregrounds and backgrounds of images of endangered animals. Then, the pre-trained image completion network CR-Fill is used to repair the incomplete environment. Furthermore, our approach identifies a foreground–background mixture of different images to produce multiple new image examples, using the relation network to permit a more realistic mixture of foreground and background images. It does not require further supervision, and it is easy to embed into existing networks, which learn to compensate for the uncertainties and nonstationarities of few-shot learning. Our experimental results are in excellent agreement with theoretical predictions by different evaluation metrics, and they unveil the future potential of video surveillance to address endangered-animal detection in studies of their behavior and conservation.

Few‐shot learning based multi‐weather‐condition impedance identification for MPPT‐controlled PV converters

The broadband impedance of converters is an essential feature for the stability analysis of new energy sources. However, obtaining the impedance for photovoltaic (PV) converters with Maximum Power Point Tracking (MPPT) control is challenging because of their non-linear control schemes and real-time changing operating points. In this case, conventional linearized impedance modelling methods are not applicable, and conventional direct measurement approaches are highly time-consuming. This paper proposes a few-shot learning approach for quick access to PV converters’ impedance. The proposed method is based on the model agnostic meta-learning (MAML) algorithm, suitable for MPPT-controlled converters whose impedance changes with time, temperature, and irradiation. In the training process, it adjusts the initial model of the machine learning algorithm under different weather conditions. After completing the training, the initial model can adapt to a new condition with very few samples. Under this approach, with only a few data measured at several frequency points, broadband impedance for MPPT-controlled converters under any weather conditions can be accurately predicted, avoiding time-consuming measurements and inaccurate prediction of existing methods. Contrast simulation results show the effectiveness and superiority of the proposed method.

An object detection-based few-shot learning approach for multimedia quality assessment

An object detection-based few-shot learning approach for multimedia quality assessment

A Few-Shot Learning Approach Assists in the Prognosis Prediction of Magnetic Resonance-Guided Focused Ultrasound for the Local Control of Bone Metastatic Lesions.

Simple SummaryWe report a local control prediction model for patients undergoing MRgFUS ablation, and provide promising guidance for clinicians to identify a suitable treatment strategy for bone metastatic lesions. We propose a few-shot learning approach to establish the quick prediction of clinical and radiographic responses. On the basis of demographic data, pre-/post-treatment immune-related cytokine change, and MRI imaging, the most suitable parameters were selected to assess potential treatment outcomes during the acute inflammatory stages within 24 h. Traditional logistic regression and few-shot learning models were compared to identify the best model on an independent test. The best predictive few-shot learning model (accuracy of 85.2%, sensitivity of 88.6%, and AUC of 0.95) was achieved by combining the clinical features with the levels of significant cytokines IL-6, IL-13, IP-10, and eotaxin.Magnetic resonance-guided focused ultrasound surgery (MRgFUS) constitutes a noninvasive treatment strategy to ablate deep-seated bone metastases. However, limited evidence suggests that, although cytokines are influenced by thermal necrosis, there is still no cytokine threshold for clinical responses. A prediction model to approximate the postablation immune status on the basis of circulating cytokine activation is thus needed. IL-6 and IP-10, which are proinflammatory cytokines, decreased significantly during the acute phase. Wound-healing cytokines such as VEGF and PDGF increased after ablation, but the increase was not statistically significant. In this phase, IL-6, IL-13, IP-10, and eotaxin expression levels diminished the ongoing inflammatory progression in the treated sites. These cytokine changes also correlated with the response rate of primary tumor control after acute periods. The few-shot learning algorithm was applied to test the correlation between cytokine levels and local control (p = 0.036). The best-fitted model included IL-6, IL-13, IP-10, and eotaxin as cytokine parameters from the few-shot selection, and had an accuracy of 85.2%, sensitivity of 88.6%, and AUC of 0.95. The acceptable usage of this model may help predict the acute-phase prognosis of a patient with painful bone metastasis who underwent local MRgFUS. The application of machine learning in bone metastasis is equivalent or better than the current logistic regression.

Meta-learning with implicit gradients in a few-shot setting for medical image segmentation

Widely used traditional supervised deep learning methods require a large number of training samples but often fail to generalize on unseen datasets. Therefore, a more general application of any trained model is quite limited for medical imaging for clinical practice. Using separately trained models for each unique lesion category or a unique patient population will require sufficiently large curated datasets, which is not practical to use in a real-world clinical set-up. Few-shot learning approaches can not only minimize the need for an enormous number of reliable ground truth labels that are labour-intensive and expensive, but can also be used to model on a dataset coming from a new population. To this end, we propose to exploit an optimization-based implicit model agnostic meta-learning (iMAML) algorithm under few-shot settings for medical image segmentation. Our approach can leverage the learned weights from diverse but small training samples to perform analysis on unseen datasets with high accuracy. We show that, unlike classical few-shot learning approaches, our method improves generalization capability. To our knowledge, this is the first work that exploits iMAML for medical image segmentation and explores the strength of the model on scenarios such as meta-training on unique and mixed instances of lesion datasets. Our quantitative results on publicly available skin and polyp datasets show that the proposed method outperforms the naive supervised baseline model and two recent few-shot segmentation approaches by large margins. In addition, our iMAML approach shows an improvement of 2%–4% in dice score compared to its counterpart MAML for most experiments.

Content-Based Video Retrieval With Prototypes of Deep Features

The rapid development in the area of information and communication technologies has enabled the transfer of high-resolution, large-sized videos, and video applications have also evolved according to data quality levels. Content-based video retrieval (CBVR) is an essential video application because it can be applied to various domains, such as surveillance, education, sports, and medicine. In this paper, we propose a CBVR method based on prototypical category approximation (PCA-CBVR), which calculates prototypes of deep features for each category to predict the user’s query video category without a classifier. We also undertake fine searching to retrieve the video most similar to the user’s query video from the predicted category database of videos. The proposed PCA-CBVR approach is efficient in terms of its computational cost and maintains meaningful information of the videos. It does not need to train a classifier even when the database is updated and uses all deep features without any dimension reduction step, such as those in CBVR studies. Moreover, we conduct fine-tuning of the 3D CNN feature extractor based on a few-shot learning approach for better domain adaptation ability and apply salient frame sampling instead of uniform frame sampling to improve the performance of the PCA-CBVR method. We demonstrate the performance capability of the proposed PCA-CBVR approach through experiments on various benchmark video datasets, in this case the UCF101, HMDB51, and ActivityNet datasets.

Defect localization during laser microwelding of battery connectors using long exposure imaging and few-shot learning

In the production of battery systems for electromobility, even a single weak electrical connection of the battery cells can lead to a failure of the entire stack. During laser welding of dissimilar materials such as aluminum and copper, real-time information about the actual part quality is essential for cost-efficient production and high product quality. To address these needs, long exposure optical imaging is used to generate detailed images during the process based on local near-infrared process emission. The high-resolution images are subsequently processed by using a novel data analysis approach. Existing image classification methods based on supervised learning usually require thousands of labeled images for training. To overcome these shortcomings, a few-shot learning approach based on Siamese Neural Networks is investigated, which require only a few samples to detect critical welding defects. The approach is evaluated on industry-relevant experiments and demonstrates promising results in terms of high defect recognition accuracies.