Generalized Category Discovery (GCD) faces the challenge of categorizing unlabeled data containing both known and novel classes, given only labels for known classes. Previous studies often treat each class independently, neglecting the inherent inter-class relations. Obtaining such inter-class relations directly presents a significant challenge in real-world scenarios. To address this issue, we propose ReLKD, an end-to-end framework that effectively exploits implicit inter-class relations and leverages this knowledge to enhance the classification of novel classes. ReLKD comprises three key modules: a target-grained module for learning discriminative representations, a coarse-grained module for capturing hierarchical class relations, and a distillation module for transferring knowledge from the coarse-grained module to refine the target-grained module’s representation learning. Extensive experiments on four datasets demonstrate the effectiveness of ReLKD, particularly in scenarios with limited labeled data. The code for ReLKD is available at https://github.com/ZhouF-ECNU/ReLKD.

Existing anomaly detection methods tend to utilize a large amount of training data to learn patterns of normal data for effective anomaly identification, but such methods typically incur substantial training time overhead. Considering that unlabeled data often contains a lot of redundant information, selecting and utilizing a small yet representative subset instead of the entire dataset can significantly improve training efficiency while maintaining detection performance. To this end, we introduce an end-to-end reinforcement learning framework with a balanced sampling strategy that targets both normal and abnormal instances. This framework identifies and exploits potential anomalies in the unlabeled data while sampling peripheral normal instances (often difficult to detect), thereby enhancing the overall anomaly detection performance without requiring excessive time for the sampling process. Additionally, we present a joint reward mechanism, combined with inconsistency penalties, which optimizes both an agent's action space and the representation space, ultimately improving the quality of the sampling process. Extensive experiments on four public datasets from different domains demonstrate the effectiveness and efficiency of our framework. The code is available at https://github.com/ZhouF-ECNU/READ.

Integrated payment platforms have significantly improved the convenience of daily life, yet they also present a fertile ground for fraudulent behavior. This paper focuses on the detection of anomalous merchants at the transaction level on such platforms, as locating specific anomalous patterns at such a granular level aids in taking corresponding security measures. However, in an integrated payment scenario, a limited number of imprecise labels are accessed at the merchant level rather than the transaction level, thus rendering transaction-level anomaly detection quite difficult. Meanwhile, the collected data comprises not only normal merchants and target anomalies (of interest) but also non-target anomalies (of lesser interest). To address these challenges, we adopt a two-step approach. First, we cluster merchants exhibiting similar behaviors and filter out potential non-target anomalies to better understand the transactional patterns among normal merchants. Then, we learn transaction representations encapsulated within hyperspheres, considering three key aspects: transaction context, historical information, and merchant information; and leverage such representations to determine anomaly scores for individual transactions. Real-world transactions from an integrated payment platform were used in the experiments. The results demonstrate that our model outperforms several state-of-the-art baselines, with an average AUPRC improvement of 10.5%-11.6%, 16.5%-16.7%, and 3.7%-5.4% in the three discovered merchant clusters.

Anomaly detection is a crucial data mining problem due to its extensive range of applications. In real-world scenarios, anomalies often exhibit different levels of priority. Unfortunately, existing methods tend to overlook this phenomenon and identify all types of anomalies into a single class. In this paper, we propose a generalized formulation of the anomaly detection problem, which covers not only the conventional anomaly detection task, but also the partial anomaly detection task that is focused on identifying target anomalies of primary interest while intentionally disregarding non-target (low-risk) anomalies. One of the challenges in addressing this problem is the overlap among normal instances and anomalies of different levels of priority, which may cause high false positive rates. Additionally, acquiring a sufficient quantity of all types of labeled non-target anomalies is not always feasible. For this purpose, we present a generalized anomaly detection framework flexible in addressing a broader range of anomaly detection scenarios. Employing a dual-center mechanism to handle relationships among normal instances, non-target anomalies, and target anomalies, the proposed framework significantly reduces the number of false positives caused by class overlap and tackles the challenge of limited amount of labeled data. Extensive experiments conducted on two publicly available datasets from different domains demonstrate the effectiveness, robustness and superior labeled data utilization of the proposed framework. When applied to a real-world application, it exhibits a lift of at least 7.08% in AUPRC compared to the alternatives, showcasing its remarkable practicality.

Anomaly detection has emerged as a prominent research area with extensive exploration across various applications. Existing methods predominantly focus on detecting all anomalies exhibiting unusual patterns, however, they overlook the critical need to prioritize the detection of target anomaly categories (anomalies of primary interest) that could pose significant threats to various systems. This oversight results in the excessive involvement of valuable human labor and resources in dealing with non-target anomalies (that are of lower interest). This work is focused on target-class anomaly detection, which entails overcoming several challenges: (1) deficient prior information regarding non-target anomalies and (2) an elevated false positive rate caused by the presence of non-target anomalies. Thus, we introduce a novel semi-supervised model, called TargAD, which leverages a few labeled target anomalies, along with potential non-target anomaly candidates and normal candidates selected from unlabeled data. By introducing a novel loss function, TargAD effectively maximizes the distributional differences among normal candidates, target anomalies, and non-target anomaly candidates, leading to a significant improvement in detecting target anomalies. Furthermore, when confronted with novel non-target anomaly scenarios, TargAD maintains its accuracy in detecting target anomalies. We conducted extensive experiments, the results of which demonstrate that TargAD outperforms eleven state-of-the-art baselines on a real-world dataset and three publicly available datasets, with average AUPRC improvements of 5.9%-24.8%, 9.2%-57.8%, 2.7%-71.3%, and 2.0%-70.3%, respectively.

Applications involving Extreme Multi-Label Classification (XMLC) face several practical challenges with respect to scale, model size and prediction latency, while maintaining satisfactory predictive accuracy. In this paper, we propose a Multi-Label Factorization Machine (MLFM) model, which addresses some of the challenges in XMLC problems. We use behavioral ad targeting as a case study to illustrate the benefits of the MLFM model. Predicting user qualifications for targeting segments plays a major role in both personalization and real-time bidding. Considering the large number of segments and the prediction time requirements of real-world production systems, building scalable models is often difficult and computationally burdensome. To cope with these challenges, we (1) reformulate the problem of assigning users to segments as a multi-label classification (XMLC) problem, and (2) leverage the benefits of the conventional FM model and generalize its capacity to joint prediction across a large number of targeting segments. We have shown that the MLFM model is both effective and computationally efficient compared to several baseline models on publicly available datasets in addition to the targeting use case.

The methods for power system event detection using field-recorded data from Phasor Measurement Units (PMUs) often require many labeled events, which can be costly or infeasible to obtain. We show that events in one power system can be accurately detected by reusing a small number of carefully selected labeled PMU data from another without the need for additional labeling. Our transfer learning-based approach outperforms alternative state-of-the-art conventional machine learning (ML) methods on a large PMU historical dataset. We demonstrate this approach with a use case of detecting events from historical PMU data recorded in the Eastern Interconnection in the USA by using similar labeled PMU data from the Western Interconnection. This technique may be propagated to other situations where some of the events’ data from one power system may be applied to enhance learning in another.

Digital media companies typically collect rich data in the form of sequences of online user activities. Such data is used in various applications, involving tasks ranging from click or conversion prediction to recommendation or user segmentation. Nonetheless, each application depends upon specialized feature engineering that requires a lot of effort and typically disregards the time-varying nature of the online user behavior. Learning time-preserving vector representations of users (user embeddings), irrespective of a specific task, would save redundant effort and potentially lead to higher embedding quality. To that end, we address the limitations of the current state-of-the-art self-supervised methods for task-independent (unsupervised) sequence embedding, and propose a novel Time-Aware Sequential Autoencoder (TASA) that accounts for the temporal aspects of sequences of activities. The generated embeddings are intended to be readily accessible for many problem formulations and seamlessly applicable to desired tasks, thus sidestepping the burden of task-driven feature engineering. The proposed TASA shows improvements over alternative self-supervised models in terms of sequence reconstruction. Moreover, the embeddings generated by TASA yield increases in predictive performance on both proprietary and public data. It also achieves comparable results to supervised approaches that are trained on individual tasks separately and require substantially more computational effort. TASA has been incorporated within a pipeline designed to provide time-aware user embeddings as a service, and the use of its embeddings exhibited lifts in conversion prediction AUC on four audiences.

OBJECTIVE: We sought to predict if patients with type 2 diabetes mellitus (DM2) would develop 10 selected complications. Accurate prediction of complications could help with more targeted measures that would prevent or slow down their development. MATERIALS AND METHODS: Experiments were conducted on the Healthcare Cost and Utilization Project State Inpatient Databases of California for the period of 2003 to 2011. Recurrent neural network (RNN) long short-term memory (LSTM) and RNN gated recurrent unit (GRU) deep learning methods were designed and compared with random forest and multilayer perceptron traditional models. Prediction accuracy of selected complications were compared on 3 settings corresponding to minimum number of hospitalizations between diabetes diagnosis and the diagnosis of complications. RESULTS: The diagnosis domain was used for experiments. The best results were achieved with RNN GRU model, followed by RNN LSTM model. The prediction accuracy achieved with RNN GRU model was between 73% (myocardial infarction) and 83% (chronic ischemic heart disease), while accuracy of traditional models was between 66% - 76%. DISCUSSION: The number of hospitalizations was an important factor for the prediction accuracy. Experiments with 4 hospitalizations achieved significantly better accuracy than with 2 hospitalizations. To achieve improved accuracy deep learning models required training on at least 1000 patients and accuracy significantly dropped if training datasets contained 500 patients. The prediction accuracy of complications decreases over time period. Considering individual complications, the best accuracy was achieved on depressive disorder and chronic ischemic heart disease. CONCLUSIONS: The RNN GRU model was the best choice for electronic medical record type of data, based on the achieved results.

OBJECTIVE: Clinical trials, prospective research studies on human participants carried out by a distributed team of clinical investigators, play a crucial role in the development of new treatments in health care. This is a complex and expensive process where investigators aim to enroll volunteers with predetermined characteristics, administer treatment(s), and collect safety and efficacy data. Therefore, choosing top-enrolling investigators is essential for efficient clinical trial execution and is 1 of the primary drivers of drug development cost. MATERIALS AND METHODS: To facilitate clinical trials optimization, we propose DeepMatch (DM), a novel approach that builds on top of advances in deep learning. DM is designed to learn from both investigator and trial-related heterogeneous data sources and rank investigators based on their expected enrollment performance on new clinical trials. RESULTS: Large-scale evaluation conducted on 2618 studies provides evidence that the proposed ranking-based framework improves the current state-of-the-art by up to 19% on ranking investigators and up to 10% on detecting top/bottom performers when recruiting investigators for new clinical trials. DISCUSSION: The extensive experimental section suggests that DM can provide substantial improvement over current industry standards in several regards: (1) the enrollment potential of the investigator list, (2) the time it takes to generate the list, and (3) data-informed decisions about new investigators. CONCLUSION: Due to the great significance of the problem at hand, related research efforts are set to shift the paradigm of how investigators are chosen for clinical trials, thereby optimizing and automating them and reducing the cost of new therapies.

Attaining the proper balance between underfitting and overfitting is one of the central challenges in machine learning. It has been approached mostly by deriving bounds on generalization risks of learning algorithms. Such bounds are, however, rarely controllable. In this study, a novel bias-variance balancing objective function is introduced in order to improve generalization performance. By utilizing distance correlation, this objective function is able to indirectly control a stability-based upper bound on a model's expected true risk. In addition, the Generalization-Aware Collaborative Ensemble Regressor (GLACER) is developed, a model that bags a crowd of structured regression models, while allowing them to collaborate in a fashion that minimizes the proposed objective function. The experimental results on both synthetic and real-world data indicate that such an objective enhances the overall model's predictive performance. When compared against a broad range of both traditional and structured regression models GLACER was ~10-56% and ~49-99% more accurate for the task of predicting housing prices and hospital readmissions, respectively.