Adversarial training (AT) has shown impressive advantages in maintaining accuracy and enhancing robustness against adversarial examples. However, most existing AT techniques jointly optimize clean-example accuracy and adversarial-example robustness as dual objectives. This setting introduces an often overlooked issue; when optimizing hard samples near the decision boundary, the model may bolster robustness at the expense of accuracy or preserve accuracy to the detriment of robustness. To alleviate the accuracy-robustness sacrifices induced by hard samples, we propose mitigating hard sample interference (MHSI) from a sample-intervention perspective. MHSI aims to reduce the instability caused by hard samples during AT. Specifically, we introduce a weighted adaptive (WA) mechanism that strengthens the model's learning of clean samples, thereby reducing the negative impact of hard samples on accuracy. In addition, guided by an analysis of the gradient norm and the Hessian matrix, we design a dynamic calibration (DC) strategy that dynamically calibrates the probability outputs of hard samples to mitigate their damage to robustness. With these two modules, our approach significantly improves robustness without sacrificing accuracy. Extensive experiments on CIFAR-10, CIFAR-100, Tiny ImageNet, and SVHN demonstrate that MHSI effectively improves both accuracy and robustness and outperforms state-of-the-art methods under glass-box attacks. Notably, under an $l_{\infty }$ attack, MHSI yields up to a 6.22% robustness gain over the AT baseline. Our code is available at https://github.com/hubin111/MHSI.

With large-scale language models demonstrating superior capabilities in a wide range of downstream natural language processing tasks, the future trajectory of research in the field of text categorization faces increasing uncertainty. In this evolving paradigm of open-ended language modeling, where task delimitations are increasingly blurred, a pressing question arises: to what extent has text classification advanced under the full potential of large language model (LLM)? To address this pivotal inquiry, we introduce recurrent generative pre-trained transformer (RGPT), an adaptive boosting framework meticulously designed to craft a dedicated LLM for text classification. RGPT constructs a sequence of base learners by dynamically modulating the training data distribution and iteratively fine-tuning LLMs. These base learners are then progressively integrated, leveraging historical prediction trajectories to form a highly specialized text classification model. Extensive empirical evaluations demonstrate that RGPT surpasses eight state-of-the-art pretrained language models and seven cutting-edge LLMs across four benchmark datasets, achieving an average performance gain of 2.90%.

In the evolving field of machine learning, ensuring group fairness has become a critical concern, prompting the development of algorithms designed to mitigate bias in decision-making processes. Group fairness refers to the principle that a model's decisions should be equitable across different groups defined by sensitive attributes such as gender or race, ensuring that individuals from privileged groups and unprivileged groups are treated fairly and receive similar outcomes. However, achieving fairness in the presence of group-specific concept drift remains an unexplored frontier, and our research represents pioneering efforts in this regard. Group-specific concept drift refers to situations where one group experiences concept drift over time, while another does not, leading to a decrease in fairness even if accuracy (ACC) remains fairly stable. Within the framework of federated learning (FL), where clients collaboratively train models, its distributed nature further amplifies these challenges since each client can experience group-specific concept drift independently while still sharing the same underlying concept, creating a complex and dynamic environment for maintaining fairness. The most significant contribution of our research is the formalization and introduction of the problem of group-specific concept drift and its distributed counterpart, shedding light on its critical importance in the field of fairness. In addition, leveraging insights from prior research, we adapt an existing distributed concept drift adaptation algorithm to tackle group-specific distributed concept drift, which uses a multimodel approach, a local group-specific drift detection mechanism, and continuous clustering of models over time. The findings from our experiments highlight the importance of addressing group-specific concept drift and its distributed counterpart to advance fairness in machine learning.

Precision medicine requires cuffless blood pressure (BP) estimation technologies to achieve high accuracy levels suitable for healthcare monitoring. Recent advances, notably in personalized BP estimation methods utilizing transfer learning, have partially met these requirements. However, traditional personalized fine-tuning pipelines often ignore key physiological factors across age groups, thus limiting their effectiveness. This study identifies an important fine-tuning principle: target subjects benefit most when fine-tuned using a source model pretrained on an age-matched group. We introduce a novel two-level ensemble learning method, ATTE, which integrates multiple BP models for cuffless BP estimation based on pulse waves. It first performs feature-level ensemble learning (FEL) to combine age-group-specific fine-tuned models and enrich representations, then applies an optimized Bayesian model averaging (BMA) method that adaptively weights model outputs to boost accuracy and reduce model-selection uncertainty. Using only 50 transfer samples, ATTE improved systolic and diastolic BP estimation accuracy by 6.55% and 5.58% on a public dataset, compared to the traditional fine-tuning paradigm. It also achieved improvements of 5.10% and 8.30% on a self-collected dataset, which has been preliminarily validated in an elderly care scenario. Ablation and interpretability analyses validated the age-matched fine-tuning principle, the effectiveness of ATTE, and its model-agnostic generalizability. This work represents a pioneering step in optimizing personalized BP estimation and underscores its significant potential for real-world healthcare applications.

Continual learning (CL) aims to empower machine learning (ML) models to learn continually from new data, while building upon previously acquired knowledge without forgetting. As models have evolved from small to large pretrained architectures, and from supporting unimodal to multimodal (MM) data, multimodal CL (MMCL) methods have recently emerged. The primary complexity of MMCL is that it extends beyond a simple stacking of unimodal CL methods. Such straightforward approaches often suffer from MM catastrophic forgetting, yielding unsatisfactory performance. In addition, MMCL introduces new challenges that unimodal CL methods fail to adequately address, including modality imbalance, complex modality interaction, high computational costs, and degradation of the pretrained zero-shot capability of MM backbones. In this work, we present the first comprehensive survey on MMCL. We provide essential background knowledge and MMCL settings, as well as a structured taxonomy of MMCL methods. We categorize MMCL methods into four categories, i.e., regularization-based, architecture-based, replay-based, and prompt-based methods, explaining their methodologies and highlighting their key innovations. Additionally, to prompt further research in this field, we summarize open MMCL datasets and benchmarks, provide an in-depth discussion, and discuss several promising future directions. We have also created a GitHub repository for indexing relevant MMCL articles and open resources available at https://github.com/LucyDYu/Awesome-Multimodal-Continual-Learning.

Dexterous grasping and manipulation with multifingered robotic hands presents a significant challenge due to their high degrees of freedom and the need for task-specific adaptations. Existing methods usually adopt single-task learning framework or focus on simple stable wrap grasping, limiting their efficiency and generalization ability when encountering new task or precise functional grasping pose. In this article, we introduce MetaGrasp, a novel approach that defines dexterous functional grasping as a multitask reinforcement learning (RL) problem based on hand grasp pose classification. Our method features a unique gradual skill curriculum learning (GSCL) framework, which structures the learning process into three stages: beginner, intermediate, and advanced curriculum learning according to the level of difficulty. MetaGrasp leverages this hierarchical learning structure to develop a versatile, adaptive grasping policy that can grasp objects based on hand grasp pose and object point cloud inputs. Taking five hand grasp types as research cases, the trained policy with our MetaGrasp can be easily adpated to grasp different object instances from different object categories according to functional grasp intentions specified by one expert demonstration without requiring extensive system interaction. We categorize the dexterous functional grasping tasks of a five-fingered robotic hand into multiple tasks based on hand poses for RL, and to combine meta imitation learning (IL) with curriculum learning. The experimental results show that the MetaGrasp has better one-shot generalization ability on new grasp tasks, and outperforms state-of-the-art single-task dexterous grasping methods.

To solve the performance degradation of visible light-based object detectors affected by low-light environments, an attention-guided detection model based on adaptive depthwise convolution is proposed. First, the quality of the original image is improved, and various image information is restored by an image enhancement module named zero-reference dual-illumination deep-curve estimation++ (Zero-DiDCE++). Then, the article further introduces a self-attention mechanism and receptive field features, designing the feature self-attention (FSA) module and the sparse receptive field adaptive depth convolution (SRDConv) module. These modules work in concert to accentuate target features and improve detection accuracy. Finally, we introduce a framework named ZRF-you only look once (YOLO)v9 for object detection, which is designed to improve performance in low-light conditions. It incorporates Zero-DiDCE++ and connects it with an improved YOLOv9 network for algorithm integration. Experiments show that based on the ExDark, Dark Face, and Tiny-Person datasets, the ZRF-YOLOv9 algorithm improves the low-light detection accuracy by 1.1%, 2.0%, and 0.5% in mAP ${}_{5}{}_{0}$ . This model not only enhances the visibility of the image but also improves detection accuracy for target images in low-light scenes.

The cooperative tracking control problem for the multiagent systems with unknown dynamic models, state constraints, control input constraints, optimal performance index, and convergence rate constraints is investigated in this article. A novel design framework is proposed to cope with these constraints and requirements. More specifically, the mean value theorem is utilized to transform the control input constraint to the unconstrained form. The performance function combined with a barrier Lyapunov function is leveraged to achieve the guaranteed transient tracking performance, finite-time convergence, and state constraints. To satisfy the optimal performance index, while considering the unknown dynamics of agents, we employ the actor-critic neural network architecture to get the near-optimal solution. Our control scheme is completely model-free. The rigorous Lyapunov stability analyses show that the full-state constraints and control input constraints are always satisfied, and the cooperative tracking errors can be made as small as possible in a finite time at a desired decay rate by appropriately setting the parameters in the performance function and barrier Lyapunov function. Finally, the simulation and hardware tests are conducted to clarify the effectiveness of the proposed control strategy.

Accurate imputation of missing data is crucial in the Industrial Internet-of-Things (IIoT), where operations are often compromised by noisy samples from harsh environments. Traditional imputation methods struggle with such noise due to their black-box nature or lack of adaptability. To address this issue, we recast data imputation as a distribution alignment challenge, utilizing the flexibility of optimal transport (OT) to handle noisy samples. Specifically, we first introduce the Proximal Optimal Transport (POT) problem, where the transportation cost is obtained by the network simplex approach with a selective matching mechanism, which renders it capable of matching distributions with noisy samples. Subsequently, we propose the POT-I framework, where the objective is to minimize the transport cost of POT. The produced gradient is used to refine the imputation value, which achieves missing data imputation (MDI) while getting robustness to noisy samples. Experiments on real-world IIoT datasets demonstrate the superiority of POT-I over state-of-the-art imputation methods.

Lightweight models are often deployed on edge devices (e.g., smartphones and wearable devices) to enhance their scalability and practicability. To enhance their generalization ability in dynamic changing environments, cloud-device collaborative learning (CDCL) is proposed to transfer the generalization ability from large models on cloud servers to lightweight models deployed on devices. However, current methods mainly focus on solving the data distribution shifts for a limited number of seen classes but ignore the continual incoming new classes. Though existing class-incremental learning (CIL) methods achieve impressive performance, two major challenges arise when adapting them into the CDCL setting: 1) poor generalization of lightweight models during CIL and 2) overfitting during data-incremental learning. In this article, we explore a new problem named class-incremental CI-CDCL, and propose a contrastive prototypical network based CI-CDCL framework, aiming to improve the effectiveness of cloud-device collaboration in both class-incremental and data-incremental learning. Extensive experiments are conducted on two public datasets and the experimental results can evaluate the effectiveness of our proposed model.