Distribution Learning Method Research Articles

A class distribution learning method for few-shot remote sensing scene classification

ABSTRACT In recent years, few-shot remote sensing scene classification has become an important foundational research topic, aiming to achieve better classification performance with limited labeled data. Due to the large intra-class variances and inter-class similarity of remote sensing scenes, the scene classification task for remote sensing is much more challenging than general few-shot image classification. In this letter, we propose a class distribution learning method under limited samples by fully using the support samples, which consists of intra-class feature aggregation (IFA) module, intra-class feature homogenization module (IFH) and inter-class cross calibration (ICC) module. IFA and IFH modules are proposed to enable the aggregation and even distribution of intra-class features around the prototypes generated by the model, reducing intra-class distances and mitigating the generation of extreme samples near decision boundaries. ICC module is designed to encourage the features generated by the model for inter-class to be further apart, thereby increasing the inter-class distance and reducing the probability of classification errors. To demonstrate the effectiveness of our proposed method, we conduct comprehensive experiments on three benchmark remote sensing scene datasets: NWPU-RESISC45, WHU-RS19, and UC Merced, achieving extremely advanced performance. Our source codes have been available at https://github.com/mitsuha97/CDLM.

Uncertainty Modeling for Gaze Estimation.

Gaze estimation is an important fundamental task in computer vision and medical research. Existing works have explored various effective paradigms and modules for precisely predicting eye gazes. However, the uncertainty for gaze estimation, e.g., input uncertainty and annotation uncertainty, have been neglected in previous research. Existing models use a deterministic function to estimate the gaze, which cannot reflect the actual situation in gaze estimation. To address this issue, we propose a probabilistic framework for gaze estimation by modeling the input uncertainty and annotation uncertainty. We first utilize probabilistic embeddings to model the input uncertainty, representing the input image as a Gaussian distribution in the embedding space. Based on the input uncertainty modeling, we give an instance-wise uncertainty estimation to measure the confidence of prediction results, which is critical in practical applications. Then, we propose a new label distribution learning method, probabilistic annotations, to model the annotation uncertainty, representing the raw hard labels as Gaussian distributions. In addition, we develop an Embedding Distribution Smoothing (EDS) module and a hard example mining method to improve the consistency between embedding distribution and label distribution. We conduct extensive experiments, demonstrating that the proposed approach achieves significant improvements over baseline and state-of-the-art methods on two widely used benchmark datasets, GazeCapture and MPIIFaceGaze, as well as our collected dataset using mobile devices.

Label distribution learning with high‐order label correlations

SummaryLabel distribution learning (LDL) is an emerging learning paradigm, which can be used to solve the label ambiguity problem. In spite of the recent great progress in LDL algorithms considering label correlations, the majority of existing methods only measure pairwise label correlations through the commonly used similarity metric, which is incapable of accurately reflecting the complex relationship between labels. To solve this problem, a novel label distribution learning method—based on high‐order label correlations (LDL‐HLC) is proposed. By virtue of the ‐regularization sparse reconstruction of the label space, the high‐order label correlations matrix is firstly obtained. Then, a new regular term can be constructed to fit the final prediction label distribution via the correction matrix. Furthermore, efficient classification performance and complete feature selection are guaranteed by common features learning via ‐regularization. Finally, the performance and effectiveness of the proposed algorithm are well illustrated through extensive experiments on 14 label distribution datasets and comparisons with some existing algorithms.

A knowledge-based approach for estimating the distribution of urban mixed land use

ABSTRACT Estimating the proportion of land-use types in different regions is essential to promote the organization of a compact city and reduce energy consumption. However, existing research in this area has a few limitations: (1) lack of consideration of land-use distribution-related factors other than POIs; (2) inability to extract complex relations from heterogeneous information; and (3) overlooking the correlation between land-use types. To overcome these limitations, we propose a knowledge-based approach for estimating land-use distributions. We designed a knowledge graph to display POIs and other related heterogeneous data and then utilized a knowledge embedding model to directly obtain the region embedding vectors by learning the complex and implicit relations present in the knowledge graph. Region embedding vectors were mapped to land-use distributions using a label distribution learning method integrating the correlation between land-use types. To prove the reliability and validity of our approach, we conducted a case study in Jinhua, China. The results indicated that the proposed model outperformed other algorithms in all evaluation indices, thus illustrating the potential of this method to achieve higher accuracy land-use distribution estimates.

Extracting statistical information about shapes in the visual environment

It is well known that observers can use so-called summary statistics of visual ensembles to simplify perceptual processing. The assumption has been that instead of representing feature distributions in detail the visual system extracts the mean and variance of visual ensembles. But recent evidence from implicit testing using a method called feature distribution learning showed that far more detail of the distributions is retained than the summary statistic literature indicates. Observers also encode higher-order statistics such as the kurtosis of feature distributions of orientation and color. But this sort of learning has not been shown for more intricate aspects of visual information. Here we tested the learning of distractor ensembles for shape, using the feature distribution learning method. Using a linearized circular shape space, we found that learning of detailed distributions of shape does not occur for this shape space while observers were able to learn the mean and range of the distributions. Previous demonstrations of feature distribution learning involved simpler feature dimensions than the more complex shape space tested here, and our findings may therefore reveal important boundary conditions of feature distribution learning.

Data-driven cooling, heating and electrical load prediction for building integrated with electric vehicles considering occupant travel behavior

The access of electric vehicles facilitates in the fluctuation and diversification of building load, accurate load prediction contributes to investigating the operation and optimization of energy supply systems for building integrated with electric vehicles (EVs). This study proposes a multivariate load prediction model for building/EVs considering occupant travel behavior. Firstly, travel variables of dissimilar occupants are obtained by fitting distribution, Markov chain and machine learning method. Besides, these variables are sampled to construct the occupant travel behavior model based on Monte Carlo method. In addition, occupancy rate and hourly load of different buildings are determined according to the occupant travel behavior. Ultimately, data-driven approaches, such as artificial neural network, long short term memory network (LSTM) and LSTM with temporal pattern attention (TPA-LSTM), are applied to construct the load prediction model for building/electric vehicles. Results indicated that building load constructed in accordance with the occupancy rate of occupants' travel behavior exhibits a favorable cooperativity with the EVs charging load. TPA-LSTM model has a high prediction accuracy, and the maximum correlation coefficient reaches 0.987. In general, this study provides an effective tool for accurate load prediction of building integrated with EVs.

Superpixel-Oriented Label Distribution Learning for Skin Lesion Segmentation.

Lesion segmentation is a critical task in skin cancer analysis and detection. When developing deep learning-based segmentation methods, we need a large number of human-annotated labels to serve as ground truth for model-supervised learning. Due to the complexity of dermatological images and the subjective differences of different dermatologists in decision-making, the labels in the segmentation target boundary region are prone to produce uncertain labels or error labels. These labels may lead to unsatisfactory performance of dermoscopy segmentation. In addition, the model trained by the errored one-hot label may be overconfident, which can lead to arbitrary prediction and model overfitting. In this paper, a superpixel-oriented label distribution learning method is proposed. The superpixels formed by the simple linear iterative cluster (SLIC) algorithm combine one-hot labels constraint and define a distance function to convert it into a soft probability distribution. Referring to the model structure of knowledge distillation, after Superpixel-oriented label distribution learning, we get soft labels with structural prior information. Then the soft labels are transferred as new knowledge to the lesion segmentation network for training. Ours method on ISIC 2018 datasets achieves an Dice coefficient reaching 84%, sensitivity 79.6%, precision 80.4%, improved by 19.3%, 8.6% and 2.5% respectively in comparison with the results of U-Net. We also evaluate our method on the tasks of skin lesion segmentation via several general neural network architectures. The experiments show that ours method improves the performance of network image segmentation and can be easily integrated into most existing deep learning architectures.

Deep Discrete Hashing for Label Distribution Learning

Label Distribution Learning (LDL) can better describe the real-world data by learning a set of label distributions instead of discrete binary labels. Particularly, hashing-based LDL has achieved promising performance due to its desirable advantages of fast similarity computation and extremely low storage cost. However, existing hashing-based LDL methods are still shallow learning methods, which cannot deeply capture the implicit data semantics, and meanwhile fail to fully model the semantic data relations. In this letter, we propose an effective and efficient Deep Discrete Hashing for Label Distribution Learning (DDH-LDL) method, which develops the first deep hashing framework for LDL. Specifically, DDH-LDL captures implicit semantic information by multi-layer non-linear transformation, and simultaneously preserves the modeled semantic relations of instances into hash codes via semantic message aggregation on Graph Convolutional Network (GCN). Furthermore, we elaborately design a discrete optimization module that is seamlessly integrated into our proposed deep hashing framework to reduce the binary quantization errors. Experiments on several widely tested datasets verify the superiority of the proposed method on both learning accuracy and efficiency.

Label Distribution Learning for Generalizable Multi-Source Person Re-Identification

Person re-identification (Re-ID) is a critical technique in the video surveillance system, which has achieved significant success in the supervised setting. However, it is difficult to directly apply the supervised model to arbitrary unseen domains due to the domain gap between the available source domains and unseen target domains. In this paper, we propose a novel label distribution learning (LDL) method to address generalizable multi-source person Re-ID task ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e</i> ., there are multiple available source domains, and the testing domain is unseen during training), which aims to explore the relation of different classes and mitigate the domain-shift across different domains so as to improve the discrimination of the model and learn the domain-invariant feature, simultaneously. Specifically, during the training process, we produce the label distribution via the online manner to mine the relation information of different classes, thus it is beneficial for extracting the discriminative feature. Besides, for the label distribution of each class, we further revise it to give more and equal attention to the other domains that the class does not belong to, which can effectively reduce the domain gap across different domains and obtain the domain-invariant feature. Furthermore, we also give the theoretical analysis to demonstrate that the proposed method can effectively deal with the domain-shift issue. Extensive experiments on multiple benchmark datasets validate the effectiveness of the proposed method and show that the proposed method can outperform the state-of-the-art methods. Besides, further analysis also reveals the superiority of the proposed method.

Co-LDL: A Co-Training-Based Label Distribution Learning Method for Tackling Label Noise

Performances of deep neural networks are prone to be degraded by label noise due to their powerful capability in fitting training data. Deeming low-loss instances as clean data is one of the most promising strategies in tackling label noise and has been widely adopted by state-of-the-art methods. However, prior works tend to drop high-loss instances directly, neglecting their valuable information. To address this issue, we propose an end-to-end framework named Co-LDL, which incorporates the low-loss sample selection strategy with label distribution learning. Specifically, we simultaneously train two deep neural networks and let them communicate useful knowledge by selecting low-loss and high-loss samples for each other. Low-loss samples are leveraged conventionally for updating network parameters. On the contrary, high-loss samples are trained in a label distribution learning manner to update network parameters and label distributions concurrently. Moreover, we propose a self-supervised module to further boost the model performance by enhancing the learned representations. Comprehensive experiments on both synthetic and real-world noisy datasets are provided to demonstrate the superiority of our Co-LDL method over state-of-the-art approaches in learning with noisy labels. The source code and models have been made available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/NUST-Machine-Intelligence-Laboratory/CoLDL</uri> .

Robust Lightweight Facial Expression Recognition Network with Label Distribution Training

This paper presents an efficiently robust facial expression recognition (FER) network, named EfficientFace, which holds much fewer parameters but more robust to the FER in the wild. Firstly, to improve the robustness of the lightweight network, a local-feature extractor and a channel-spatial modulator are designed, in which the depthwise convolution is employed. As a result, the network is aware of local and global-salient facial features. Then, considering the fact that most emotions occur as combinations, mixtures, or compounds of the basic emotions, we introduce a simple but efficient label distribution learning (LDL) method as a novel training strategy. Experiments conducted on realistic occlusion and pose variation datasets demonstrate that the proposed EfficientFace is robust under occlusion and pose variation conditions. Moreover, the proposed method achieves state-of-the-art results on RAF-DB, CAER-S, and AffectNet-7 datasets with accuracies of 88.36%, 85.87%, and 63.70%, respectively, and a comparable result on the AffectNet-8 dataset with an accuracy of 59.89%. The code is public available at https://github.com/zengqunzhao/EfficientFace.

Emotional Attention Detection and Correlation Exploration for Image Emotion Distribution Learning

Current works on image emotion distribution learning typically extract visual representations from the holistic image or explore emotion-related regions in the image from a global-wise perspective. However, different regions of an image contribute differently to the arousal of each emotion. Existing works do not deeply explore corresponding emotion-aware regions of each emotion in the image, nor do they fully capture the relationship between each emotion-aware region and the emotion labels. In this article, we propose a novel attention based emotion distribution learning method, which can explore the emotion-related regions of images from the perspective of each emotion category, and can conduct region relationship learning. Specifically, we introduce a semantic guided attention detection network to generate class-wise attention maps for each emotion and a global-wise attention map for the holistic image. Meanwhile, an emotional graph-based network is adopted to capture the correlation between each region and the emotion distribution. Experiments on several benchmark datasets demonstrate the superiority of the proposed method compared to related works.

Deep Conditional Distribution Learning for Age Estimation

Age estimation is a challenging task not only because face appearance is affected by illumination, pose, and expression, but also because there exists age label ambiguity among different demographic groups. In this work, we first revisit different label distribution learning (LDL) based age estimation methods and propose a more general formulation, which can unify individual LDL-based age estimation methods, as well as the traditional regression, classification, and ranking based age estimation methods. Based on such a general formulation, we propose a novel deep conditional distribution learning (DCDL) method, which can flexibly leverage a varying number of auxiliary face attributes to achieve adaptive age-related feature learning and improve age estimation robustness against the challenges above. Experimental results on multiple age estimation datasets (MORPH II, AgeDB, FG-NET, MegaAge-Asian, CLAP2016, UTK-Face, and LFW+) show that the proposed approach outperforms the state-of-the-art age estimation methods by a large margin. In addition, the proposed approach can generalize well to other human attributes estimation tasks, like height, weight, and body mass index (BMI) estimation.

Retracted] Age Label Distribution Learning Based on Unsupervised Comparisons of Faces

Although label distribution learning has made significant progress in the field of face age estimation, unsupervised learning has not been widely adopted and is still an important and challenging task. In this work, we propose an unsupervised contrastive label distribution learning method (UCLD) for facial age estimation. This method is helpful to extract semantic and meaningful information of raw faces with preserving high‐order correlation between adjacent ages. Similar to the processing method of wireless sensor network, we designed the ConAge network with the contrast learning method. As a result, our model maximizes the similarity of positive samples by data enhancement and simultaneously pushes the clusters of negative samples apart. Compared to state‐of‐the‐art methods, we achieve compelling results on the widely used benchmark, i.e., MORPH.

Day‐ahead renewable scenario forecasts based on generative adversarial networks

With the increasing penetration of renewable resources such as wind and solar, especially in terms of large-scale integration, the operation and planning of power systems are faced with great risks due to the inherent stochasticity of natural resources. Although this uncertainty can be anticipated, the timing, magnitude, and duration of fluctuations cannot be predicted accurately. In addition, the outputs of renewable power sources are correlated in space and time, and this brings further challenges for predicting the characteristics of their future behavior. To address these issues, this paper describes an unsupervised distribution learning method for renewable scenario forecasts that considers spatiotemporal correlation based on generative adversarial network (GAN), which has been shown to generate realistic time series for stochastic processes. We first utilize an improved GAN to learn unknown data distributions and model the dynamic processes of renewable resources. We then generate a large number of forecasted scenarios using stochastic constrained optimization. For validation, we use power generation data from the National Renewable Energy Laboratory wind and solar integration datasets. The simulation results show that the generated trajectories not only reflect the future power generation dynamics, but also correctly capture the temporal, spatial, and fluctuant characteristics of the real power generation processes. The experimental comparisons verify the superiority of the proposed method and indicate that it can reduce at least 50% of the training iterations of the generative model for scenario forecasts.

Facial Depression Recognition by Deep Joint Label Distribution and Metric Learning

While existing prediction models built on popular deep architectures have shown promising results in facial depression recognition, they still lack sufficient discriminative power due to the issues of 1) limited amount of labeled depression data for deep representation learning and, 2) large variation in facial expression across different persons of the same depression score and the subtle difference in facial expression across different depression levels. In this article, we formulate the facial depression recognition as a label distribution learning (LDL) problem, and propose a deep joint label distribution and metric learning (DJ-LDML) method to address these issues. In DJ-LDML, LDL exploits label relevance inherent in depression data to implicitly increase the amount of training data associated with each depression level without actually enlarging the dataset, while deep metric learning (DML) aims at learning a deep ordinal embedding with a specifically designed label-aware histogram loss, allowing semantics similarity between video sequences (described by ordinal labels) to be preserved for discriminative feature learning. The two learning modules in our DJ-LDML work collaboratively to enhance the representation ability and discriminative power of the deeply learned spatiotemporal feature, leading to improved depression prediction. We empirically evaluate our method on two benchmark datasets and the results demonstrate the effectiveness of our formulation.

Quantum Generative Adversarial Networks for learning and loading random distributions

Quantum algorithms have the potential to outperform their classical counterparts in a variety of tasks. The realization of the advantage often requires the ability to load classical data efficiently into quantum states. However, the best known methods require {mathcal{O}}left({2}^{n}right) gates to load an exact representation of a generic data structure into an n-qubit state. This scaling can easily predominate the complexity of a quantum algorithm and, thereby, impair potential quantum advantage. Our work presents a hybrid quantum-classical algorithm for efficient, approximate quantum state loading. More precisely, we use quantum Generative Adversarial Networks (qGANs) to facilitate efficient learning and loading of generic probability distributions - implicitly given by data samples - into quantum states. Through the interplay of a quantum channel, such as a variational quantum circuit, and a classical neural network, the qGAN can learn a representation of the probability distribution underlying the data samples and load it into a quantum state. The loading requires {mathcal{O}}left(polyleft(nright)right) gates and can thus enable the use of potentially advantageous quantum algorithms, such as Quantum Amplitude Estimation. We implement the qGAN distribution learning and loading method with Qiskit and test it using a quantum simulation as well as actual quantum processors provided by the IBM Q Experience. Furthermore, we employ quantum simulation to demonstrate the use of the trained quantum channel in a quantum finance application.

Incomplete label distribution learning based on supervised neighborhood information

Label distribution learning (LDL) assumes labels are associated with each instance to some degree and tries to model the relationship between labels and instances. LDL has achieved great success in many applications, but most existing LDL methods are designed for data with complete annotation information. However, in reality, supervised information often be incomplete due to the huge costs of data collection. In this paper, we propose a novel incomplete label distribution learning method based on supervised neighborhood information (IncomLDL-SNI). The proposed method uses partial least squares to project the original data into a supervised feature space where instances with similar labels are likely to be projected together. Then, IncomLDL-SNI utilizes the Euclidean distance to find the nearest neighbors for target samples in the supervised feature space and recovers the missing annotations from the neighborhood label Information. Extensive experiments on various data sets validate the effectiveness of our proposal.

Filling missing values by local reconstruction for incomplete label distribution learning

Label distribution learning (LDL) deals with the problems when we care more about the relative importance of different labels in the description of an instance, where labels are associated with each instance to some degree. LDL has achieved great success in many applications, but most existing LDL methods cannot handle learning tasks with incomplete annotation information. In this paper, we propose a novel incomplete label distribution learning method based on local reconstruction (IncomLDL-LR). Both the feature matrix and label information are included in a unified Principal Component Analysis (PCA) model to describe the intrinsic structure of original data in the supervised low-dimensional space. Based on the reasonable assumption that the incomplete label of each instance can be linearly reconstructed from its neighbours' labels, IncomLDL-LR gradually recovers the missing label values by the averaged column score of corresponding neighbours in the PCA space. The proposed algorithms are compared with state-of-the-art algorithms using five LDL evaluation metrics on 15 public datasets. Extensive experiments validate the effectiveness of our proposal.

Filling missing values by local reconstruction for incomplete label distribution learning

Label distribution learning (LDL) deals with the problems when we care more about the relative importance of different labels in the description of an instance, where labels are associated with each instance to some degree. LDL has achieved great success in many applications, but most existing LDL methods cannot handle learning tasks with incomplete annotation information. In this paper, we propose a novel incomplete label distribution learning method based on local reconstruction (IncomLDL-LR). Both the feature matrix and label information are included in a unified Principal Component Analysis (PCA) model to describe the intrinsic structure of original data in the supervised low-dimensional space. Based on the reasonable assumption that the incomplete label of each instance can be linearly reconstructed from its neighbours' labels, IncomLDL-LR gradually recovers the missing label values by the averaged column score of corresponding neighbours in the PCA space. The proposed algorithms are compared with state-of-the-art algorithms using five LDL evaluation metrics on 15 public datasets. Extensive experiments validate the effectiveness of our proposal.