The emergence of multimodal technologies has propelled Vision-Language Incremental Learning (VLIL) into a research spotlight. Current VLIL approaches predominantly inherit unimodal paradigms, failing to address fundamental distinctions between visual and linguistic modalities. Crucially, the semantic gap between images and text creates divergent learning dynamics: visual data exhibits rich, distributed information while textual representations remain explicit and compact. Consequently, textual elements align with class-specific tasks, whereas individual images inherently span multiple such tasks, creating dual bottlenecks in class-level memory allocation and scene-level knowledge transfer. To overcome these challenges, we propose DCIM (Dual Class-Individual Memory), a novel framework featuring complementary mechanisms for vision-language continual learning. For class-level constraints, our Hierarchical Class Memory Management (HCMM) strategy dynamically allocates memory resources across object categories. It employs forgetting simulation to identify and preserve the most vulnerable samples, ensuring robust long-term knowledge retention. For scene-level adaptation, the Scene Reconstruction Memory(SRM) module captures generalized environmental representations, enabling contextual transfer to novel classes and disambiguation of semantically related concepts within shared scenes.Extensive experiments on two vision-language tasks, i.e., visual question answering (VQA) and Image captioning (IC), demonstrate the effectiveness and excellent generalization ability of our approach, achieving superior performance under continual learning settings.

While imitation learning (IL) emerges as a promising paradigm for embodied intelligent robots, its practical application is constrained by slow execution speeds, caused by the computational intensity of precise multi-model trajectory prediction, especially in complex dynamic environments. In contrast, humans can efficiently perform long-duration tasks through subconscious-driven habitual actions, such as riding bikes, without focusing on execution details. Motivated by this insight, we proposed Subconscious Robotic Imitation Learning (SRIL) framework, which mimicked the subconscious information extraction and decision-making abilities through intent-aware sampling and cognitive hierarchical reasoning, thereby significantly improving IL task execution efficiency. Experimental results demonstrated that execution speeds of the SRIL were 100% to 200% faster over SOTA policies for comprehensive bimanual tasks, with consistently higher success rates.

This paper proposed M3DSS, a multi-platform, multi-sensor, and multi-scenario dataset for Simultaneous Localization and Mapping (SLAM) systems. Fifty-five sequences were collected from multiple platforms, including a handheld equipment, an unmanned ground vehicle, a quadruped robot, a car, and an unmanned aerial vehicle. Sensors used in M3DSS included two pairs of stereo event cameras with resolutions of$640\times 480$and$346\times 260$, one infrared camera, four RGB cameras, two visual-inertial sensors, four mechanical and one solid-state LiDARs, three inertial measurement units, two global navigation satellite and inertial navigation systems with real-time kinematic signals. 21 various sensors were used on 5 different platforms under various challenging scenarios, including extreme illumination, aggressive motion, low-texture, high-speed driving scenarios, etc. To the best of our knowledge, M3DSS offered the richest event-based sensory information for SLAM up to date. We comprehensively evaluated state-of-the-art SLAM approaches and identified their limitations on M3DSS. Details could be found at https://neufs-ma.github.io/M3DSS.

Behavior cloning (BC) is a widely used method for learning from expert demonstrations due to its simplicity and efficiency. However, the reliability and stability of BC decline when facing data distribution shifts, especially in single-arm robots with limited fields of view. This study introduces a Geometrically and Historically Constrained Behavior Cloning (GHCBC) method, where an HCBC module utilizes visual and action histories to capture temporal dependencies, maximizing the use of available information, and a GCBC module incorporates high-level perceptual data, such as the relative poses of joints and end-effectors, to enhance BC performance. Experiments demonstrate that the GHCBC outperforms current SOTA BC methods, achieving a 31.5% improvement in simulation success rates and 48.4% in real-robot scenarios respectively. To the best of our knowledge, this is the first time that the GHCBC has been introduced in robotic BC where great potential is demonstrated for long-term tasks in real world environments.

A microgrid is a localized distribution network that consists of electricity users who have access to local renewable and other energy sources. Typically, it is connected to a utility distribution grid but can also function independently. The utility grid plays a crucial role in the nation's economy and security, as well as the well-being of its residents. However, connecting different microgrids to a wider network through a utility's substation can expose them to significant cyber threats. This study examines cybersecurity vulnerabilities in microgrid systems using a testbed approach. By simulating various cyberattack scenarios on a microgrid testbed, we assess the impact of different attack types on microgrid operations, with a particular focus on system stability and communication networks. Attacks such as denial of service, communication hijacking, and others are explored. Our findings highlight significant weaknesses in the communication infrastructure and provide insights into designing microgrids that can effectively address cybersecurity challenges in real-world industrial utility networks.

The rapid advancement of deep learning and time-frequency analysis techniques have brought about a revolution in fault diagnosis for mechanical systems, offering flexible and efficient solutions. Nevertheless, data imbalance issues continue to pose significant obstacles in fault diagnosis modeling. In this research, we propose the use of a domain adaptation generative adversarial network (DAGAN) that capitalizes on transfer learning to extract valuable information from the majority-class data, while concurrently generating and augmenting minority-class data to expand the training dataset. DAGAN incorporates advanced techniques, including deep domain confusion and parameter forgetting, to enhance knowledge extraction and transfer during the transfer learning process, resulting in more realistic and comprehensive generation outcomes when dealing with small sample training. Furthermore, we have developed an imbalanced fault diagnosis method based on DAGAN, which further incorporates Continuous Wavelet Transform and Deep Residual Networks. Finally, the effectiveness and superiority of proposed method are validated on bearing and gearbox datasets. The experimental results demonstrate the outstanding performance of our method in effectively addressing imbalanced fault diagnosis. Note to Practitioners—In this paper, we present a practical approach to enhance the diagnosis of imbalanced faults. Our approach begins by utilizing Time-frequency images, which offer a comprehensive representation of the temporal and spectral characteristics of mechanical systems’ behavior. These images serve as input features and form the foundation for our fault diagnosis modeling. To address the limitations imposed by imbalanced data, we introduce DAGAN and incorporate advanced techniques such as deep domain confusion and parameter forgetting. These techniques facilitate the extraction and transfer of knowledge during the transfer learning process of DAGAN. Consequently, our approach generates more realistic and comprehensive outcomes, even when confronted with limited training samples. To validate the effectiveness and superiority of our proposed approach, we conducted extensive experiments on bearing and gearbox datasets. The results of these experiments demonstrate that our practical approach, which combines wavelet transform-based time-frequency analysis and the innovative DAGAN framework, offers a reliable and comprehensive solution for overcoming challenges associated with imbalanced fault data.

Visible and infrared image fusion(VIF) provides more comprehensive understanding of a scene and can facilitate subsequent processing. Although frequency domain contains valuable global information in low frequency and rapid pixel intensity variation data in high frequency of images, existing fusion methods mainly focus on spatial domain. To close this gap, a novel VIF method in frequency domain is proposed. First, a frequency-domain feature extraction module is developed for source images. Then, a frequency-domain transformer fusion method is designed to merge the extracted features. Finally, a residual reconstruction module is introduced to obtain final fused images. To the best of our knowledge, it is the first time that image fusion study is conducted from frequency domain perspective. Comprehensive experiments on three datasets, i.e., MSRS, TNO, and Roadscene, demonstrate that the proposed approach obtains superior fusion performance over several state-of-the-art fusion methods, indicating its great potential as a generic backbone for VIF tasks.

Industrial intelligent systems (IIS) play a huge role in modern industry, and their intelligent models of IIS enable diagnosis of faults, key performance indicator (KPI) prediction, and other important industrial process analysis in a data-driven way. However, the performance of intelligent models is limited by the quantity and quality of local data in specific factories. At the same time, the privacy information and security concerns contained by industrial data lead to the problem of information silos in industry. This hinders data sharing and cross-factory collaborations. To address these issues, this article makes the following contributions. First, for the first time, we empower industrial federated big data analytics (IFBDA) of IIS with swarm learning, and propose a hyperledger fabric-based IFBDA blockchain (IFBDAchain) for multifactory information sharing and collaborative modeling. Second, in the IFBDAchain, we further consider potential dishonest behaviors among federated members, and design verification and privacy protection mechanisms to ensure trustworthiness of analytics. Third, we validate the IFBDAchain with two real industrial cases. The results demonstrate the effectiveness of the IFBDAchain in fault classification and KPI prediction tasks in industry. Compared to the average values of local learning, our method increases the classification accuracy by 27.6%, 69.4%, and 33.1% under Independent and identically distributed (IID), non-IID, and unbalanced conditions, respectively. Furthermore, the root-mean-square error of the KPI prediction decreases by 33.3%, 49%, and 45.7% for the IID, non-IID, and unbalanced conditions, respectively, indicating its significant potential as a generic backbone for industrial federated Big Data analytics.

This work-in-progress paper introduces the AIca-demic system, an innovative framework employing Agentic AI and Agile methodologies to enhance learning in complex domains such as computer networks. Positioned within the topic of AI and Machine Learning Tools to Enhance Instruction, it aims to revolutionize the learning experience and outcomes for the intricate subject matter. This system emphasizes active, adaptive learning experiences through AI generated or AI improved educational materials with multiple iterations of feedback and improvement cycles. Utilizing fine tuned large language models (LLM), the AIcademic system assembles an interactive AI team, e.g. AIcademic Professor, Student, and Instructional Designer. Each AI agent is uniquely configured with our POISE prompt engineering model to analyze and simulate real-time classroom interactions from multiple viewpoints. Active learning pedagogy is embedded into the system through prompt engineering during the creation of each agent. Agile methodology is employed to organize collaborations of the AI agents for complex task planning and implementation, feedback integration, output con-tinuous improvement, and agent self-enhancement. A suite of AI tools is explored to dynamically create tailored educational materials aligned with the educator's teaching preferences and students' needs. Preliminary results from a pilot implementation of teaching the transport layer in computer networks demon-strated improvements in student engagement and comprehension over previous materials. This AIcademic framework presents a promising and scalable paradigm for AI applications in educational environments. While still under development, this research aims to refine and expand these findings, exploring the full potential of integrating Prompt Engineering and Agentic AI for creating active learning environments across complex technical subjects. The implications extend beyond computer network education, offering a blueprint to redefine teaching and learning in a technology-enhanced era. We invite collaboration from the broader academic community to refine the Agent prompt design, automate AI to AI interactions, assess long term impacts, and explore further applications.

As computer networking technologies such as Ethernet gain momentum in modern industrial control systems (ICSs), deterministic delay performance, which refers to the worst-case latencies provisioned by the network infrastructure, has become a critical property required by real-time control, automation, and operations. Although analytical methods that identify exact worst-case bounds on flow-specific communication delays have been proposed, there remain two hurdles discouraging further explorations of network-wide deterministic delay performance: (i) methods based on mixed integer linear programming (MILP) have high computational complexities; and (ii) existing methods are flow-specific. This paper proposes a deep-learning-assisted approach to understanding the deterministic delay performance of networked industrial control systems. Transforming flow-specific worst-case delay bounding into network-wide deterministic delay analysis, our approach facilitates the incorporation of application-specific hard-real-time performance constraints. By incorporating graph neural networks (GNNs), our approach significantly reduces the runtime costs and provides sufficiently tight approximations to the optimal MILP solutions. Through a combination of numerical analyses and simulations, we demon-strate that our approach enables the timely exploration of various worst-case scenarios, thereby paving the way for agile service provisioning in time-critical and/or delay-sensitive ICSs as well as model-based design of self-healing networked ICSs.

Unmanned aerial vehicle (UAV) is becoming a crucial aerial platform to provide emergency or enhanced communication and sensing services benefited from its unique features, including agile mobility and high probability of Line-of-Sight coverage. In this article, we investigate the novel UAV trajectory design problem where the UAV communicates with multiple users and simultaneously senses the positions of multiple targets, via the integrated communication and sensing design. To evaluate the overall system utility, we first derive the communication throughput and localization error Cramér-Rao bound (CRB) model and highlight the coupled challenge and tradeoffs on UAV’s trajectory. We thus model the three typical integrated sensing and communication scenarios: 1) communication centric; 2) sensing centric; and 3) tradeoff scenarios. We further introduce path discretization to support in-flight communication and hovering sensing, which make it possible to jointly optimize the UAV trajectory, communication throughput and target localization estimation error CRB with limited complexity. Because this optimization problem involves integer programming caused by multiuser association, we propose a novel trajectory initialization framework based on traveling salesman problem to determine the service order for sensing targets and the initial UAV trajectory. To address the nonconvexity of this optimization problem, we propose an efficient iterative algorithm using the successive convexity approximation technique to obtain the approximate optimal solution, which is dynamically reconfigured and optimized with updated information in flight. Extensive numerical results demonstrate that the proposed algorithm achieves superior performance in the tradeoff between average achievable rate and CRB in all three scenarios.

The joint use of multisource remote-sensing data for Earth observation has drawn much attention due to its robust performance. Although many methods have been proposed to fuse multimodal data, they tend to improve the interaction of different modality data while ignoring the optimization of each modality. Existing studies show that high-performance modalities will suppress the learning of weak ones, leading to under-optimized multimodal learning. To this end, we propose a general framework called gradient decoupled network (GDNet) to assist the multimodal remote sensing (RS) classification. GDNet guides each modality encoder in the multimodal model to learn probabilistic representations instead of deterministic ones. This helps decouple their gradient, reducing their influence on each other and encouraging them to learn the modality-specific information. Then, we further introduce the unimodal regularization for each modality encoder to align their logit output with the multimodal one and label distribution simultaneously. This helps introduce independent gradient paths for each morality encoder to accelerate their optimization when preserving the modality-share information. Finally, extensive experiments conducted on three benchmark datasets demonstrate that the proposed GDNet can effectively address the under-optimized problem in multimodal RS image classification. Code is available athttps://github.com/shicaiwei123/TGRS-GDNet.

Multi-agent Deep Reinforcement Learning (MADRL) based traffic signal control lbecomes a popular research topic in recent years. To alleviate the scalability issue of completely centralized reinforcement learning (RL) techniques and the non-stationarity issue of completely decentralized RL techniques on large-scale traffic networks, some literature utilizes a regional control approach where the whole network is firstly partitioned into multiple disjoint regions, followed by applying the centralized RL approach to each region. However, the existing partitioning rules either have no constraints on the topology of regions or require the same topology for all regions. Meanwhile, no existing regional control approach explores the performance of optimal joint action in an exponentially growing regional action space when intersections are controlled by 4-phase traffic signals (EW, EWL, NS, NSL). In this paper, we propose a novel RL training framework named RegionLight to tackle the above limitations. Specifically, the topology of regions is firstly constrained to a star network which comprises one center and an arbitrary number of leaves. Next, the network partitioning problem is modeled as an optimization problem to minimize the number of regions. Then, an Adaptive Branching Dueling Q-Network (ABDQ) model is proposed to decompose the regional control task into several joint signal control sub-tasks corresponding to particular intersections. Subsequently, these sub-tasks maximize the regional benefits cooperatively. Finally, the global control strategy for the whole network is obtained by concatenating the optimal joint actions of all regions. Experimental results demonstrate the superiority of our proposed framework over all baselines under both real and synthetic scenarios in all evaluation metrics.

SCRUM (Scrum) is a popular framework for developing and sustaining complex products. It is designed to be an agile methodology that emphasizes collaboration, flexibility, and customer satisfaction. At its core, Scrum is built around the concept of a small, self-organizing team of people, known as a Scrum Team. This team works together to achieve a common goal, each member taking on a specific role and responsibility. In this paper, we propose a modified Scrum Team structure as a model for student teams in academic project environments. Our modification involves identifying specific roles within the team to provide more structure and align with industry practices. By doing this, we hope to help students better understand the importance of roles and responsibilities in project management and improve their project outcomes. The proposed approach involves providing a Scrum mapping for student teams of three and four members, with each member taking on a specific role based on Scrum principles. In our proposed approach, roles will rotate after each new assignment or a set number of times, giving students the opportunity to experience different roles and responsibilities within the team. We believe this approach will provide further exposure to the popular agile design approach and help students develop critical project management skills. We implemented our modified approach in a computer engineering course with lab components focused on microcontrollers to test our modified approach. Our experience using this approach was positive, with students reporting a greater understanding of project management concepts and improved project outcomes. Finally, we believe that our modified Scrum Team structure will also aid in the assessment of ABET's student outcome 5, which focuses on teamwork, communication, and project management. By providing a clear framework for project management and emphasizing the importance of roles and responsibilities within the team, we hope to help students develop the skills they need to succeed in their future careers.

Learning based multimodal data has attracted increasing interest in the remote sensing community owing to its robust performance. Although it is preferable to collect multiple modalities for training, not all of them are available in practical scenarios due to the restriction of imaging conditions. Therefore, how to assist the model inference with missing modalities is significant for multimodal remote sensing image processing. In this work, we propose a general framework called modality-shared hallucination network (MSH-Net) to address this issue by reconstructing complete modality-shared features from the incomplete inference modalities. Compared to conventional privilege modality hallucination methods, MSH-Net does not only help preserve the cross-modal interactions for model inference, but also scales well with the increasing number of missing modalities. We further develop a novel joint adaptation distillation (JAD) method that guides the hallucination model to learn the modality-shared knowledge from the multimodal model by matching the joint probability distributions between representation and groundtruth. This overcomes the representation heterogeneity caused by the discrepancy between inputs and structures of multimodal and hallucination model, while preserving the decision boundaries refined by multimodal cues. Finally, extensive experiments conducted on four common modality combinations demonstrate that the proposed MSH-Net can effectively address the problem of missing modalities and achieve state-of-the-art performance. Code is available at: https://github.com/shicaiwei123/MSHNet.

Without any knowledge of the vehicle model and its parameters, a novel complete model-free path-following control strategy is developed for autonomous vehicles based on the time-delay estimation (TDE) technique. Different from the existing time-delay control (TDC) approaches, an adaptive nonsingular terminal sliding mode (ANTSM) control law is designed to stabilize the path-following errors without any information of the suitable control gain, which is significant in the conventional TDC scheme, and the boundary of the TDE error, which is necessary for the sliding-mode-based control scheme. The proposed model-free control structure can dynamically update the gain of the designed controller and the boundaries of the TDE error, and the practical finite-time convergence of the preview error can be achieved. In the HIL tests, the comparative results demonstrate that the proposed ANTSM model-free control strategy can provide superior comprehensive tracking performance over both the model-based sliding mode controller and the conventional TDC controller, while the autonomous vehicle follows desired paths.

Guqin is a plucked seven-string traditional Chinese musical instrument that exists for over 3,000 years. However, as an Intangible World Cultural Heritage, the inheritance of Guqin and its culture in modern society is in deep danger. According to our study with 1,006 Chinese worldwide, Guqin as an instrument is not well-known and barely accessible. To better promote Guqin, we developed two interactive systems: VirGuqin and MRGuqin. VirGuqin was developed using a low-cost motion tracking device and was tested in a museum. 89% of 308 participants expressed an increase in interest in learning Guqin after using our system. MRGuqin was developed as a mixed reality learning environment to reduce the entry barrier to Guqin, and was tested by 16 participants, allowing them to learn Guqin significantly faster and perform better than the current practice. Our study demonstrates how technology can be used to help the inheritance of this dying art.

Online teaching has imposed great challenges for student engagement during the pandemic. Building a virtual classroom with active student participation is our approach to address some of the concerns and make online learning more effective. In this paper, Microsoft Teams is introduced as a virtual study room with many virtual tables (channels), where each table (channel) serves as an integrated platform for group meetings. Within MS Teams, Learner-learner interaction is boosted by virtual meetings, group poster boards, the “mention” function, and emojis. By integrating MS Teams with Zoom meetings, we can offer a zero blackout, fully interactive learning environment. The paper includes a detailed description of the required technologies for such a delivery, time requirements for the design and delivery of such an approach, and faculty assessment and perspective of the methodology. Finally, a summary of the advantages, disadvantages, and student feedback is included in the paper.

Since the problem of optimal radio channel assignment (RCA) is NP-hard, the existing RCA algorithms have to use heuristic approaches for any networks of practical sizes. However, the algorithms suffer from two major issues, i.e., unable to consider the co-channel interference (CCI), and finding only sub-optimal assignment. In this work, we propose a new strategy to overcome the two challenging issues. (1) To consider CCI among neighboring head routers (HRs) and their terminals, we propose to choose the average effective channel utilization (ECU) of an HR as the basic objective function. The function can be also the sum of the average ECU, for which the optimization target is directly to maximize the overall throughput. (2) We propose to use the simulated annealing (SA) algorithm to find the optimal assignment and use the Gibbs sampling (GS) technique to convert the global optimization problem to a series of local optimization problems. In this way, we propose a distributed optimal assignment, i.e., SA-GS-based RCA (SRCA) algorithm. Our extensive simulation results have demonstrated that SRCA outperforms all the existing RCA algorithms in achieving global optimality with bounded CS.

For IEEE 802.11 wireless LANs (WLAN) with multiple access points (AP), it is critical to allocate the limited number of radio channels dynamically and efficiently. This paper proposes a new framework for radio channel allocation (RCA) strategy for WLANs to support multimedia traffic. Two techniques are introduced for this purpose: service differentiation in the MAC layer, and link adaptation in the physical layer. First, we formulate the RCA as a min-max optimization problem regarding channel utilization with constraints of transmitting power and data rates. Second, we derive an expression to evaluate the channel utilization, which incorporates the condition of a wireless channel, such as signal-to-noise (SNR) ratio and transmitting power, in addition to the transmitting probability of a station. It also incorporates the differentiations among the stations within an AP on behalf the traffic flows carried by the links between a station and its AP. Third, we propose a new RCA algorithm that considers both link adaptation and service differentiation mechanisms, which have not been considered by the existing RCA schemes. Simulation results have demonstrated the effectiveness of the proposed strategy.