Abstracts:Kleinberg (2002) introduced three axioms to formalize the behavior of clustering algorithms and presented that no clustering algorithm satisfies them. However, this result usually is inconsistent with the practical experience of clustering algorithms. In this article, we reformulate these axioms to fill the gap and verify the existence of clustering algorithms satisfying the modified axioms when the point set is well-clusterable. In particular, the concept of $(\delta ,\varepsilon)$ –K segmentation is proposed to characterize the set that has the potential to be clustered well. Then, we verify the existence and the uniqueness of $(\delta ,\varepsilon)$ –K segmentation of a set, respectively. Next, we demonstrate that the $(\delta ,\varepsilon)$ –K segmentation is compatible with K–means, Min-Cut, DBSCAN, and several clustering internal evaluation (CIE) indexes. In addition, the ratio $\delta / \varepsilon $ can be treated as the measure of not only characterizing well-clusterable sets but also of evaluating the performance of clustering results, respectively.

Abstracts:Support vector machines (SVMs) have been the learning model of choice in numerous classification applications. While SVMs are widely successful in real-world deployments, they remain susceptible to mislabeled examples in training datasets where the presence of few faults can severely affect decision boundaries, thereby affecting the model’s performance on unseen data. In this brief, we develop and describe in implementation detail a novel method based on $L_{1}$ -norm principal-component data analysis and geometry that aims to filter out atypical data instances on a class-by-class basis before the training phase of SVMs and thus provide the classifier with robust support-vector candidates for making classification boundaries. The proposed dataset curation method is entirely data-driven (touch-free), unsupervised, and computationally efficient. Extensive experimental studies on real datasets included in this brief illustrate the $L_{1}$ -norm curation method and demonstrate its efficacy in protecting SVM models from data faults during learning.

Abstracts:The success of large-scale pretrained models has established fine-tuning as a standard method for achieving significant improvements in downstream tasks. However, fine-tuning the entire parameter set of a pretrained model is costly. Parameter-efficient transfer learning (PETL) has recently emerged as a cost-effective alternative for adapting pretrained models to downstream tasks. Despite its advantages, the increasing model size and input resolution present challenges for PETL, as the training memory consumption is not reduced as effectively as the parameter usage. In this article, we introduce fine-grained prompt tuning plus (FPT+), a PETL method designed for high-resolution medical image classification, which significantly reduces the training memory consumption compared to other PETL methods. FPT+ performs transfer learning by training a lightweight side network and accessing pretrained knowledge from a large pretrained model (LPM) through fine-grained prompts and fusion modules. Specifically, we freeze the LPM of interest and construct a learnable lightweight side network. The frozen LPM processes high-resolution images to extract fine-grained features, while the side network employs corresponding downsampled low-resolution images to minimize memory usage. To enable the side network to leverage pretrained knowledge, we propose fine-grained prompts and fusion modules, which collaborate to summarize information through the LPM’s intermediate activations. We evaluate FPT+ on eight medical image datasets of varying sizes, modalities, and complexities. Experimental results demonstrate that FPT+ outperforms other PETL methods, using only 1.03% of the learnable parameters and 3.18% of the memory required for fine-tuning an entire ViT-B model. Our code is available https://github.com/YijinHuang/FPT.

Abstracts:The deep reinforcement learning (DRL) has made significant progress in various simulation environments. However, applying DRL methods to real-world scenarios poses certain challenges due to limitations in visual fidelity, scene complexity, and task diversity within existing environments. To address limitations and explore the potential ability of DRL, we developed a 3-D open-world first-person shooter (FPS) game called Unreal BattleGround (UBG) using the unreal engine (UE). UBG provides a realistic 3-D environment with variable complexity, random scenes, diverse tasks, and multiple scene interaction methods. This benchmark involves far more complex state-action spaces than classic pseudo-3-D FPS games (e.g., ViZDoom), making it challenging for DRL to learn human-level decision sequences. Then, we propose the object-aware hierarchically proximal policy optimization (OaH-PPO) method in the UBG. It involves a two-level hierarchy, where the high-level controller is tasked with learning option control, and the low-level workers focus on mastering subtasks. To boost the learning of subtasks, we propose three modules: an object-aware module for extracting depth detection information from the environment, potential-based intrinsic reward shaping for efficient exploration, and annealing imitation learning (IL) to guide the initialization. Experimental results have demonstrated the broad applicability of the UBG and the effectiveness of the OaH-PPO. We will release the code of the UBG and OaH-PPO after publication.

Abstracts:Semantic change detection (CD) not only helps pinpoint the locations where changes occur, but also identifies the specific types of changes in land cover and land use. Currently, the mainstream approach for semantic CD (SCD) decomposes the task into semantic segmentation (SS) and CD tasks. Although these methods have achieved good results, they do not consider the incentive effect of task correlation on the entire model. Given this issue, this article further elucidates the SCD task through the lens of multitask learning theory and proposes a semantic change detection network based on boundary detection and task interaction (BT-SCD). In BT-SCD, the boundary detection (BD) task is introduced to enhance the correlation between the SS task and the CD task in SCD, thereby promoting positive reinforcement between SS and CD tasks. Furthermore, to enhance the communication of information between the SS and CD tasks, the pixel-level interaction strategy and the logit-level interaction strategy are proposed. Finally, to fully capture the temporal change information of the bitemporal features and eliminate their temporal dependency, a bidirectional change feature extraction module is proposed. Extensive experimental results on three commonly used datasets and a nonagriculturalization dataset (NAFZ) show that our BT-SCD achieves state-of-the-art performance. The code is available at https://github.com/TangYJ1229/BT-SCD

Abstracts:Multiplayer game theory has been widely studied, with most existing research focusing on fully connected network structures. In contrast, multiplayer graphical games consider sparser communication topologies, making them more practical for large-scale systems. This article, based on a reinforcement learning (RL) method, investigates the problem of computing Nash equilibrium (NE) strategies in a class of multiplayer graphical games where the system is influenced by an external system. To estimate the unknown states of the external system, we propose a distributed adaptive observer and prove that its observation error asymptotically converges to zero. Furthermore, we derive a range of discount factor values that preserve system stability. To solve for the NE strategy, we develop an off-policy algorithm integrated with the distributed adaptive observer for policy evaluation. To enhance convergence speed, we introduce a distributed policy improvement mechanism, which ensures policy convergence to equilibrium while maintaining system stability. The effectiveness of the proposed algorithm is validated through simulations on a voltage synchronization system.

Abstracts:Data-driven methods have significantly advanced optimal learning control, but some approaches overlook systematic considerations of data utilization, including safety, efficiency, and error accumulation. To address the neglects in safe neural critic control, this article introduces a parallel multistep evaluation mechanism that combines data from the system interaction with data generated by data-driven models. Based on this evaluation mechanism, we propose a novel parallel multistep Q-learning algorithm that enhances data utilization efficiency and mitigates the error accumulation. Furthermore, we formulate a novel control barrier function (CBF) to ensure safety during learning and control processes, which is capable of dealing with asymmetric constraints and adjusting the constraint strength. In addition, the analysis reveals that multistep information introduced by data-driven models influences the learning performance of actor–critic neural networks (NNs). Finally, parallel multistep Q-learning, which makes use of data in aspects of safety, efficiency, and error bounds, is validated within an orbital maneuver system.

Abstracts:The dynamics of 2-D two-grid coupled cellular neural networks (CNNs) are considered. Assuming the boundary conditions of zero-flux type, the linearized model is analyzed by using the decoupling method, which is described as matrix operations, such as the Kronecker product and Kronecker sum. Then, the local stability of the zero equilibrium related to system parameters is studied. Based on the results, the sufficient conditions that induce Turing instability are derived. Furthermore, as a special kind of Turing patterns, the occurrence conditions for Hopf bifurcations are considered. In addition, the global stability of the zero equilibrium is analyzed by constructing proper Lyapunov functions. Finally, numerical simulations are given to illustrate the theoretical results.

Abstracts:Self-supervised contrastive learning (SCL) has driven significant advancements in time-series representation learning. While recent studies built upon the information noise contrastive estimation (InfoNCE) loss framework focus on constructing appropriate positives and negatives, we theoretically analyze and identify two overlooked issues inherent in this approach: false negatives and class imbalance. To address these challenges, we propose a simple yet effective modification based on the SimCLR framework, integrating a multi-instance discrimination task to mitigate false negatives. Additionally, we introduce a graph-based interactive projection head and semantic consistency regularization, which enhances minority-class representations with minimal annotation cost. Extensive experiments on six real-world time-series datasets demonstrate that our approach consistently outperforms state-of-the-art methods, achieving up to 3.96% higher accuracy and 10.73% improvement in $F1$ -score, particularly benefiting imbalanced data scenarios.

Abstracts:The federated learning (FL) paradigm can significantly solve the rising public concern about data privacy in affective computing. However, conventional FL methods perform poorly due to the uniqueness of the task, as the personalized emotion data vary from client to client. To resolve the privacy-utility paradox, this work proposes a framework that largely improves federated affective computing (FAC) via calibrating the global feature space and communicating privacy-agnostic auxiliary information. The framework consists of two components: first, an emotion hemisphere (EH) representation structure is proposed, which utilizes emotional prior knowledge to unify the emotion global feature space of different clients. Second, the server uses the normalized parameter importance matrix to guide the model aggregation. It retains crucial parameters for individual local models, thereby alleviating the slow convergence problem in the global model caused by the skewed label distribution. The proposed framework yields significant performance gains, and extensive experiments on three emotion datasets demonstrate the effectiveness and the practicality of our approach.