Candidate Labels Research Articles

Partial label metric learning by collapsing classes

Partial label learning (PLL) is a weakly supervised learning framework proposed recently, in which the ground-truth label of training sample is not precisely annotated but concealed in a set of candidate labels, which makes the accuracy of the existing PLL algorithms is usually lower than that of the traditional supervised learning algorithms. Since the accuracy of a learning algorithm is usually closely related to its distance metric, the metric learning technologies can be employed to improve the accuracy of the existing PLL algorithms. However, only a few PLL metric learning algorithms have been proposed up to the present. In view of this, a novel PLL metric learning algorithm is proposed by using the collapsing classes model in this paper. The basic idea is first to take each training sample and its neighbor with shared candidate labels as a similar pair, while each training sample and its neighbor without shared candidate labels as a dissimilar pair, then two probability distributions are defined based on the distance and label similarity of these pairs, respectively, finally the metric matrix is obtained via minimizing the Kullback–Leibler divergence of these two probability distributions. Experimental results on six UCI data sets and four real-world PLL data sets show that the proposed algorithm can obviously improve the accuracy of the existing PLL algorithms.

Partial Multi-Label Learning via Credible Label Elicitation.

Partial multi-label learning (PML) deals with the problem where each training example is associated with an overcomplete set of candidate labels, among which only some candidate labels are valid. The task of PML naturally arises in learning scenarios with inaccurate supervision, and the goal is to induce a multi-label predictor which can assign a set of proper labels for unseen instance. The PML training procedure is prone to be misled by false positive labels concealed in the candidate label set, which serves as the major modeling difficulty for partial multi-label learning. In this paper, a novel two-stage PML approach is proposed which works by eliciting credible labels from the candidate label set for model induction. In the first stage, the labeling confidence of candidate label for each PML training example is estimated via iterative label propagation. In the second stage, by utilizing credible labels with high labeling confidence, multi-label predictor is induced via pairwise label ranking coupled with virtual label splitting or maximum a posteriori (MAP) reasoning. Experimental studies show that the proposed approach can achieve highly competitive generalization performance by excluding most false positive labels from the training procedure via credible label elicitation.

Partial Multi-Label Learning with Label Distribution

Partial multi-label learning (PML) aims to learn from training examples each associated with a set of candidate labels, among which only a subset are valid for the training example. The common strategy to induce predictive model is trying to disambiguate the candidate label set, such as identifying the ground-truth label via utilizing the confidence of each candidate label or estimating the noisy labels in the candidate label sets. Nonetheless, these strategies ignore considering the essential label distribution corresponding to each instance since the label distribution is not explicitly available in the training set. In this paper, a new partial multi-label learning strategy named Pml-ld is proposed to learn from partial multi-label examples via label enhancement. Specifically, label distributions are recovered by leveraging the topological information of the feature space and the correlations among the labels. After that, a multi-class predictive model is learned by fitting a regularized multi-output regressor with the recovered label distributions. Experimental results on synthetic as well as real-world datasets clearly validate the effectiveness of Pml-ld for solving PML problems.

Deep Discriminative CNN with Temporal Ensembling for Ambiguously-Labeled Image Classification

In this paper, we study the problem of image classification where training images are ambiguously annotated with multiple candidate labels, among which only one is correct but is not accessible during the training phase. Due to the adopted non-deep framework and improper disambiguation strategies, traditional approaches are usually short of the representation ability and discrimination ability, so their performances are still to be improved. To remedy these two shortcomings, this paper proposes a novel approach termed “Deep Discriminative CNN” (D2CNN) with temporal ensembling. Specifically, to improve the representation ability, we innovatively employ the deep convolutional neural networks for ambiguously-labeled image classification, in which the well-known ResNet is adopted as our backbone. To enhance the discrimination ability, we design an entropy-based regularizer to maximize the margin between the potentially correct label and the unlikely ones of each image. In addition, we utilize the temporally assembled predictions of different epochs to guide the training process so that the latent groundtruth label can be confidently highlighted. This is much superior to the traditional disambiguation operations which treat all candidate labels equally and identify the hidden groundtruth label via some heuristic ways. Thorough experimental results on multiple datasets firmly demonstrate the effectiveness of our proposed D2CNN when compared with other existing state-of-the-art approaches.

HERA

Partial label learning (PLL) aims to learn from the data where each training instance is associated with a set of candidate labels, among which only one is correct. Most existing methods deal with this type of problem by either treating each candidate label equally or identifying the ground-truth label iteratively. In this article, we propose a novel PLL approach named HERA, which simultaneously incorporates the HeterogEneous Loss and the SpaRse and Low-rAnk procedure to estimate the labeling confidence for each instance while training the desired model. Specifically, the heterogeneous loss integrates the strengths of both the pairwise ranking loss and the pointwise reconstruction loss to provide informative label ranking and reconstruction information for label identification, whereas the embedded sparse and low-rank scheme constrains the sparsity of ground-truth label matrix and the low rank of noise label matrix to explore the global label relevance among the whole training data, for improving the learning model. Comprehensive ablation study demonstrates the effectiveness of our employed heterogeneous loss, and extensive experiments on both artificial and real-world datasets demonstrate that our method achieves superior or comparable performance against state-of-the-art methods.

Partial Label Learning with Batch Label Correction

Partial label (PL) learning tackles the problem where each training instance is associated with a set of candidate labels, among which only one is the true label. In this paper, we propose a simple but effective batch-based partial label learning algorithm named PL-BLC, which tackles the partial label learning problem with batch-wise label correction (BLC). PL-BLC dynamically corrects the label confidence matrix of each training batch based on the current prediction network, and adopts a MixUp data augmentation scheme to enhance the underlying true labels against the redundant noisy labels. In addition, it introduces a teacher model through a consistency cost to ensure the stability of the batch-based prediction network update. Extensive experiments are conducted on synthesized and real-world partial label learning datasets, while the proposed approach demonstrates the state-of-the-art performance for partial label learning.

Partial Multi-Label Learning with Noisy Label Identification

Partial multi-label learning (PML) deals with problems where each instance is assigned with a candidate label set, which contains multiple relevant labels and some noisy labels. Recent studies usually solve PML problems with the disambiguation strategy, which recovers ground-truth labels from the candidate label set by simply assuming that the noisy labels are generated randomly. In real applications, however, noisy labels are usually caused by some ambiguous contents of the example. Based on this observation, we propose a partial multi-label learning approach to simultaneously recover the ground-truth information and identify the noisy labels. The two objectives are formalized in a unified framework with trace norm and ℓ1 norm regularizers. Under the supervision of the observed noise-corrupted label matrix, the multi-label classifier and noisy label identifier are jointly optimized by incorporating the label correlation exploitation and feature-induced noise model. Extensive experiments on synthetic as well as real-world data sets validate the effectiveness of the proposed approach.

Multi-View Partial Multi-Label Learning with Graph-Based Disambiguation

In multi-view multi-label learning (MVML), each training example is represented by different feature vectors and associated with multiple labels simultaneously. Nonetheless, the labeling quality of training examples is tend to be affected by annotation noises. In this paper, the problem of multi-view partial multi-label learning (MVPML) is studied, where the set of associated labels are assumed to be candidate ones and only partially valid. To solve the MVPML problem, a two-stage graph-based disambiguation approach is proposed. Firstly, the ground-truth labels of each training example are estimated by disambiguating the candidate labels with fused similarity graph. After that, the predictive model for each label is learned from embedding features generated from disambiguation-guided clustering analysis. Extensive experimental studies clearly validate the effectiveness of the proposed approach in solving the MVPML problem.

Explaining a bag of words with hierarchical conceptual labels

In natural language processing and information retrieval tasks, the bag-of-words model is widely used to represent the semantics of texts. However, it is difficult for machines to sufficiently understand a bag of words as well as the corresponding text without explicit semantic explanation, thus hindering the power of the bag-of-words model in many scenarios. In this paper, we introduce the task of hierarchical conceptual labeling (HCL), which aims to generate a set of conceptual labels with a hierarchy to explicitly explain the semantics of a bag of words, where the candidate labels are selected from a large-scale knowledge base, i.e., Microsoft Concept Graph. To this end, we first propose a denoising algorithm to filter out the noise in a bag of words in advance. Then the hierarchical conceptual labels are generated for the clean bag of words based on a hierarchical clustering algorithm, i.e., Bayesian rose trees. We conduct extensive experiments and prove that (1) the proposed denoising algorithm can effectively delete the noise words from a bag of words, (2) the Bayesian rose trees based algorithm can generate hierarchical conceptual labels for a bag of words with a high accuracy.

Open Set Audio Recognition for Multi-Class Classification With Rejection

Most supervised audio recognition systems developed to this point have used a testing set which includes the same categories as the training set database. Such systems are called closed-set recognition (CSR). However, audio recognition in real applications can be more complicated, where the datasets can be dynamic, and novel categories can ceaselessly be detected. Hence, in practice, the usual methods will assign to these novel classes labels which are often incorrect. This work aims to investigate audio open-set recognition (OSR) suitable for multi-classes classification recognition, with a rejection option for classes never seen by the system. A probabilistic calibration of a support vector machine classifier is utilized and formulated under the open-set scenario. For this, it is proposed to apply a threshold technique called peak side ratio (PSR) to the audio recognition task. A candidate label is first examined by a Platt-calibrated support vector machine (SVM) to produce posterior probabilities. The PSR is then used to characterize the distribution of posterior probabilities values. This process helps to determine a threshold in order to reject or accept a particular class. Our proposed method is evaluated on different variations of open sets, using well-known metrics. Experimental results reveal that our proposed method outperforms previous OSR approaches over a wide range of openness values.

Imprecise Deep Forest for Partial Label Learning

In partial label (PL) learning, each instance corresponds to a set of candidate labels, among which only one is valid. The objective of PL learning is to obtain a multi-class classifier from the training instances. Because the true label of a PL training instance is hidden in the candidate label set and inaccessible to the learning algorithm, the training process of the classifier is significantly challenging. This study proposes a novel deep learning method for PL learning based on the improved error-correcting output codes (ECOC) algorithm and deep forest (DF) framework. For the ECOC algorithm, we extract the prior knowledge of the candidate label sets from the PL training set to optimize the generation of its coding matrix, where different binary training sets can be derived from the PL training set based on the dichotomy corresponding to each column code. For the DF framework, this improved ECOC algorithm is embedded as a unit in its cascade structure; moreover, an imprecise evaluation method is designed to determine the growth of the cascade of the DF. The effectiveness of the proposed method is verified by conducting several experiments on artificial and real-world PL datasets.

CrossPatch-Based Rolling Label Expansion for Dense Stereo Matching

We present a novel algorithm called crosspatch-based rolling label expansion for accurate stereo matching. This optimization-based approach can effectively estimate the 3D label of each pixel from huge and infinite label space and then generate a continuous disparity map. The algorithm has two obvious characteristics when compared with the traditional label expansion algorithms. The first feature is the cross-based multilayer structure, where each layer contains a series of cross patches with adaptive shapes, reflecting the edge structure of objects on the image. Besides, such cross patches are non-overlapping and independent, satisfying the submodular property for employing graph cuts. The second feature is the rolling optimization, that firstly generates new label proposal by expanding candidate labels within cross patches, then globally updates labels for the whole image using a proposed rolling move. The experimental results show the high matching accuracy of our method, both in pixel level and subpixel level. According to the latest ranking list of Middlebury 3.0 benchmark, our method is one of the best stereo matching algorithms.

Groupwise Ranking Loss for Multi-Label Learning

This work studies multi-label learning (MLL), where each instance is associated with a subset of positive labels. For each instance, a good multi-label predictor should encourage the predicted positive labels to be close to its ground-truth positive ones. In this work, we propose a new loss, named Groupwise Ranking LosS (GRLS) for multi-label learning. Minimizing GRLS encourages the predicted relevancy scores of the ground-truth positive labels to be higher than that of the negative ones. More importantly, its time complexity is linear with respect to the number of candidate labels, rather than square complexity for some pairwise ranking based methods. We further analyze GRLS in the perspective of label-wise margin and suggest that multi-label predictor is label-wise effective if and only if GRLS is optimal. We also analyze the relations between GRLS and some widely used loss functions for MLL. Finally, we apply GRLS to multi-label learning, and extensive experiments on several benchmark multi-label databases demonstrate the competitive performance of the proposed method to state-of-the-art methods.

A Self-Ensemble Approach for Partial Multi-Label Learning

Partial multi-label learning (PML), which tackles the problem where each training instance is associated with multiple candidate labels which only a subset are valid. In this paper, we propose a simple but effective batch-wise PML model, PML-SE, which tackles PML problem with a self-ensemble approach (SE), namely the ensembles of model and predictions. Specially, PML-SE introduces a teacher model to refine a more reliable soft label matrix of each training batch by iteratively ensembling the current learned prediction network with the formal one in an online manner. Besides, it adopts a MixUp data augmentation scheme to enhance the robustness of the prediction network against the redundant irrelevant labels. In addition, we form self-ensemble label predictions through a consistency cost to boost the performance of the prediction network. Extensive experiments are conducted on synthesized and real-world PML datasets, while the proposed approach demonstrates the state-of-the-art performance for partial multi-label learning.

Large Margin Partial Label Machine.

Partial label learning (PLL) is a multi-class weakly supervised learning problem where each training instance is associated with a set of candidate labels but only one label is the ground truth. The main challenge of PLL is how to deal with the label ambiguities. Among various disambiguation techniques, large margin (LM)-based algorithms attract much attention due to their powerful discriminative performance. However, existing LM-based algorithms either neglect some potential candidate labels in constructing the margin or introduce auxiliary estimation of class capacities which is generally inaccurate. As a result, their generalization performances are deteriorated. To address the above-mentioned drawbacks, motivated by the optimistic superset loss, we propose an LM Partial LAbel machiNE (LM-PLANE) by extending multi-class support vector machines (SVM) to PLL. Compared with existing LM-based disambiguation algorithms, LM-PLANE considers the margin of all potential candidate labels without auxiliary estimation of class capacities. Furthermore, an efficient cutting plane (CP) method is developed to train LM-PLANE in the dual space. Theoretical insights into the effectiveness and convergence of our CP method are also presented. Extensive experiments on various PLL tasks demonstrate the superiority of LM-PLANE over existing LM based and other representative PLL algorithms in terms of classification accuracy.

Taking a machine’s perspective: Humans can decipher adversarial images

How similar is the human visual system to the sophisticated machine-learning systems that mirror its performance? Models of object categorization based on convolutional neural networks (CNNs) have achieved human-level benchmarks in labeling novel images. These advances not only support new technologies, but may also serve as candidate models for human vision itself. However, unlike human vision, CNNs can be “fooled” by adversarial examples — carefully crafted images that appear as nonsense patterns to humans but are recognized as familiar objects by machines, or that appear as one object to humans and a different object to machines. This extreme divergence between human and machine classification challenges the promise of these new advances, both as applied image-recognition systems and as models of human vision. Surprisingly, however, little work has empirically investigated human classification of adversarial stimuli; do humans and machines fundamentally diverge? Here, we show that human and machine classification of adversarial stimuli are robustly related. We introduce a “machine-theory-of-mind” task in which observers are shown adversarial images and must anticipate the machine’s label from a set of various alternatives. Across eight experiments on five prominent and diverse adversarial imagesets, human subjects reliably identified the machine’s preferred labels over relevant foils. We observed this result not only in forced-choice settings between two candidate labels, but also when subjects freely chose among dozens of possible labels. Moreover, this pattern persisted for images with strong antecedent identities (e.g., an orange adversarially perturbed into a “power drill”), and even for images described in the literature as “totally unrecognizable to human eyes” (e.g., unsegmented patterns of colorful pixels that are classified as an “armadillo”). We suggest that human intuition may be a more reliable guide to machine (mis)classification than has typically been imagined, and we explore the consequences of this result for minds and machines alike.

SmartLoc: Smart Wireless Indoor Localization Empowered by Machine Learning

Recently, machine learning (ML) has been widely adopted for fingerprint-based indoor localization because of its potency in delineating relationships between received signal strength (RSS) information and labels accurately. Existing ML-based indoor localization systems are less robust because they only adopt the output with the highest probability. This affects the final location estimate, hence compromising accuracy due to the severity of RSS fluctuations. Since different ML algorithms (MLAs) yield different performances, it is therefore intuitive to fuse predictions from multiple MLAs to improve the positioning performance in the presence of signal fluctuation. In this article, we propose SmartLoc, a smart wireless indoor localization framework to enhance indoor localization. In the offline phase, multiple MLAs are trained by utilizing an offline database. We further apply probability alignment to guarantee the predicted probabilities of each MLA at the same confidence level. In the online phase, given a testing RSS sample of a user at an unknown location, we extract the labels with probabilities greater than a certain threshold from each MLA to construct the space of candidate labels (SCL). The size of SCL can be adaptively determined by using our proposed dynamic size determination algorithm. Based on the SCL, we propose a probabilistic model to intelligently estimate the user's location by evaluating the label credibility simultaneously. A high label credibility indicates that the frequently occurred label is more likely to be true. Experimental results in a real changing environment verify the superiority of SmartLoc, outperforming the best among comparative methods by 10.8% in 75th percentile accuracy.

Disambiguation-free partial label learning

Partial label learning is an important weakly supervised machine learning framework. In partial label learning, each object is described by a single instance in the input space; however, in the output space, it is associated with a set of candidate labels among which only one is valid. An intuitive strategy is to disambiguate candidate labels, but this strategy tends to be misled by false positive labels; therefore, new disambiguation-free approaches need to be considered. In this paper, several algorithms are reviewed from the perspective of disambiguation and disambiguation-free strategies. First, the problem definition on partial label learning and its relationship with other related learning frameworks are given. Second, several representative partial label learning algorithms via the disambiguation strategy are introduced. Third, two of our proposed disambiguation-free algorithms are presented. Finally, the summary of this paper is given and possible future investigations on partial label learning are briefly discussed.

Partial label learning via low-rank representation and label propagation

In partial label learning, each training instance is assigned with a set of candidate labels, among which only one is correct. An intuitive strategy to learn from such ambiguous data is disambiguation. Existing methods following such strategy either identify the ground-truth label via treating each candidate label equally or disambiguate the candidate label set via assuming latent variable and optimizing it iteratively. In this paper, we propose a novel two-stage method called partial label learning via low-rank representation and label propagation, where instance similarity and label confidence are taken into consideration to improve the disambiguation ability of the model. In the first stage, we first build the global instance–similarity relationship via low rank representation and sparse constraint and then obtain the accurate instance-label assignments via iterative label propagation strategy. In the second stage, we utilize the Adaboost.R2 to make prediction for unseen instances, where CART is incorporated as the weak classifier. Extensive experiments on the artificial and real-world data sets demonstrate that the proposed method can achieve superior or comparable performance than the comparing state-of-the-art approaches.

Partial Label Learning via Label Enhancement

Partial label learning aims to learn from training examples each associated with a set of candidate labels, among which only one label is valid for the training example. The common strategy to induce predictive model is trying to disambiguate the candidate label set, such as disambiguation by identifying the ground-truth label iteratively or disambiguation by treating each candidate label equally. Nonetheless, these strategies ignore considering the generalized label distribution corresponding to each instance since the generalized label distribution is not explicitly available in the training set. In this paper, a new partial label learning strategy named PL-LE is proposed to learn from partial label examples via label enhancement. Specifically, the generalized label distributions are recovered by leveraging the topological information of the feature space. After that, a multi-class predictive model is learned by fitting a regularized multi-output regressor with the generalized label distributions. Extensive experiments show that PL-LE performs favorably against state-ofthe-art partial label learning approaches.