Multi-objective multi-armed bandit (MO-MAB) problems traditionally aim to achieve Pareto optimality. However, real-world scenarios often involve users with varying preferences across objectives, resulting in a Pareto-optimal arm that may score high for one user but perform quite poorly for another. This highlights the need for customized learning, a factor often overlooked in prior research. To address this, we study a preference-aware MO-MAB framework in the presence of explicit user preference. It shifts the focus from achieving Pareto optimality to further optimizing within the Pareto front under preference-centric customization. To our knowledge, this is the first theoretical study of customized MO-MAB optimization with explicit user preferences. Motivated by practical applications, we explore two scenarios: unknown preference and hidden preference, each presenting unique challenges for algorithm design and analysis. At the core of our algorithms are preference estimation and preference-aware optimization mechanisms to adapt to user preferences effectively. We further develop novel analytical techniques to establish near-optimal regret of the proposed algorithms. Strong empirical performance confirm the effectiveness of our approach.

Micronuclei detection is crucial in clinical diagnostics and environmental toxicology. Despite advances in computer vision and AI, challenges persist due to the significant scale difference between large cell images and small micronucleus targets. We propose a methon based on re-parameterization multi-layer adaptive feature fusion (YOLO-HDM), an innovative approach that integrates multi-scale feature extraction with re-parameterized multi-layer adaptive feature fusion. Our multi-scale feature extraction module (HDC) enhances the backbone network by preserving spatial information during downsampling and adapting to irregular input shapes. To improve small target detection accuracy while reducing computational demands, we introduce RepGFPN-HM, a lossless re-parameterization generalized feature pyramid network. By combining this with MAFF, our Rep-MAFF method mitigates feature inconsistencies across detection layers and enhances multi-scale target detection. Experimental results demonstrate YOLO-HDM achieves 96.1% mAP50 on the cell micronucleus dataset with 94.3% recall and 60.0% mAP50-95. When tested on a public tuberculosis dataset, it outperforms the state-of-the-art YOLOv5-CTS model by 2%, validating both its effectiveness and generalization capability. The code is available at: https://github.com/Aaclf/YOLO-HDM

Automated seizure detection (ASD) using intracranial electroencephalography (iEEG) is critical for effective epilepsy treatment. However, the significant domain shift of iEEG signals across subjects poses a major challenge, limiting their applicability in real-world clinical scenarios. In this paper, we address this issue by analyzing the primary cause behind the failure of existing iEEG models for subject-independent seizure detection, and identify a critical universal seizure pattern: seizure events consistently exhibit higher average amplitude compared to adjacent normal events. To mitigate the domain shifts and preserve the universal seizure patterns, we propose a novel self-comparison mechanism. This mechanism effectively aligns iEEG signals across subjects and time intervals. Building upon these findings, we propose Difference Matrix-based Neural Network (DMNet), a subject-independent seizure detection model, which leverages self-comparison based on two constructed (contextual, channel-level) references to mitigate shifts of iEEG, and utilize a simple yet effective difference matrix to encode the universal seizure patterns. Extensive experiments show that DMNet significantly outperforms previous SOTAs while maintaining high efficiency on a real-world clinical dataset collected by us and two public datasets for subject-independent seizure detection. Moreover, the visualization results demonstrate that the generated difference matrix can effectively capture the seizure activity changes during the seizure evolution process. Additionally, we deploy our method in an online diagnosis system to illustrate its effectiveness in real clinical applications.

Due to the ubiquity of graph data on the web, web graph mining has become a hot research spot. Nonetheless, the prevalence of largescale web graphs in real applications poses significant challenges to storage, computational capacity and graph model design. Despite numerous studies to enhance the scalability of graph models, a noticeable gap remains between academic research and practical web graph mining applications. One major cause is that in most industrial scenarios, only a small part of nodes in a web graph are actually required to be analyzed, where we term these nodes as target nodes, while others as background nodes. In this paper, we argue that properly fetching and condensing the background nodes from massive web graph data might be a more economical shortcut to tackle the obstacles fundamentally. To this end, we make the first attempt to study the problem of massive background nodes compression for target nodes classification. Through extensive experiments, we reveal two critical roles played by the background nodes in target node classification: enhancing structural connectivity between target nodes, and feature correlation with target nodes. Following this, we propose a novel Graph-Skeleton model, which properly fetches the background nodes, and further condenses the semantic and topological information of background nodes within similar target-background local structures. Extensive experiments on various web graph datasets demonstrate the effectiveness and efficiency of the proposed method. In particular, for MAG240M dataset with 0.24 billion nodes, our generated skeleton graph achieves highly comparable performance while only containing 1.8% nodes of the original graph.

Recently, deep learning has been proven to be a promising approach in standard dynamic range (SDR) image compression. However, due to the wide luminance distribution of high dynamic range (HDR) images and the lack of large standard datasets, developing a deep model for HDR image compression is much more challenging. To tackle this issue, we view HDR data as distributional shifts of SDR data and the HDR image compression can be modeled as an out-of-distribution generalization (OoD) problem. Herein, we propose a novel out-of-distribution (OoD) HDR image compression framework (OoDHDR-codec). It learns the general representation across HDR and SDR environments, and allows the model to be trained effectively using a large set of SDR datases supplemented with much fewer HDR samples. Specifically, OoDHDR-codec consists of two branches to process the data from two environments. The SDR branch is a standard blackbox network. For the HDR branch, we develop a hybrid system that models luminance masking and tone mapping with white-box modules and performs content compression with black-box neural networks. To improve the generalization from SDR training data on HDR data, we introduce an invariance regularization term to learn the common representation for both SDR and HDR compression. Extensive experimental results show that the OoDHDR codec achieves strong competitive in-distribution performance and state-of-the-art OoD performance. To the best of our knowledge, our proposed approach is the first work to model HDR compression as OoD generalization problems and our OoD generalization algorithmic framework can be applied to any deep compression model in addition to the network architectural choice demonstrated in the paper. Code available at https://github.com/caolinfeng/OoDHDR-codec.

Graph neural networks (GNNs) have been intensively studied in various real-world tasks. However, the homophily assumption of GNNs' aggregation function limits their representation learning ability in heterophily graphs. In this paper, we shed light on the path level patterns in graphs that can explicitly reflect rich semantic and structural information. We therefore propose a novel Structure-aware Path Aggregation Graph Neural Network (PathNet) aiming to generalize GNNs for both homophily and heterophily graphs. Specifically, we first introduce a maximal entropy path sampler, which helps us sample a number of paths containing structural context. Then, we introduce a structure-aware recurrent cell consisting of order-preserving and distance-aware components to learn the semantic information of neighborhoods. Finally, we model the preference of different paths to target node after path encoding. Experimental results demonstrate that our model achieves superior performance in node classification on both heterophily and homophily graphs.

It is ambitious to develop a brain-controlled robotic arm for some patients with motor impairments to perform activities of daily living using brain–computer interfaces (BCIs). Despite much progress achieved, this mission is still very challenging mainly due to the poor decoding performance of BCIs. The problem is even exacerbated in the case of noninvasive BCIs. A shared control strategy is developed in this work to realize flexible robotic arm control for reach and grasp of multiple objects. With the intelligent assistance provided by robot vision, the subject was only required to finish gross reaching movement and target selection using a simple motor imagery-based BCI with binary output. Along with the user control, the robotic arm, which identified and localized potential targets within the workspace in the background, was capable of providing both trajectory correction in the reaching phase to reduce trajectory redundancy and autonomous grasping assistance in the phase of grasp. Ten subjects participated in the experiments containing one session of two-block grasping tasks with fixed locations and another one of randomly placed three-block grasping tasks. The results of the experiments demonstrated substantial improvement with the shared control system. Compared with the single BCI control, the success rate of shared control was significantly higher ($p < 0.001$for group performance), and moreover, the task completion time and perceived difficulty were significantly lower ($p < 0.001$for group performance both), indicating the potential of our proposed shared control system in real applications.Note to Practitioners—This article is motivated by the problem of dexterous robotic arm control based on a brain–computer interface (BCI). For people suffering from severe neuromuscular disorders or accident injuries, a brain-controlled robotic arm is expected to provide assistance in their daily lives. A primary bottleneck to achieve the objective is that the information transfer rate of current BCIs is not high enough to produce multiple and reliable commands during the online robotic control. In this work, machine autonomy is incorporated in a BCI-controlled robotic arm system, where the user and machine can work together to reach and grasp multiple objects in a given task. The intelligent robot system autonomously localized the potential targets and provided trajectory correction and grasping assistance accordingly. Meanwhile, the user only needed to complete gross reaching movement and target selection with a basic binary motor imagery-based BCI, which reduced the task difficulty and retained the volitional involvement of the user at the same time. The results of the experiments showed that the accuracy and efficiency of grasping tasks increased significantly in the shared control mode together with a significant decrease in the perceived mental workload, which indicates that our proposed shared control system is effective and user-friendly in practice. In the future, more feedback information will be introduced to enhance the task performance further, and a wheelchair-mounted robotic arm system will be developed for greater flexibility. In addition, more functional task modules (e.g., self-feeding and opening doors) should be integrated for more practical utilities.