Search Results (18,453)

Large language models (LLMs) remain vulnerable to jailbreak attacks, especially in black-box settings where target-model gradients and internal tokenization are inaccessible. Recent information bottleneck-based defenses cast prompt protection as a compression problem, but existing methods still rely heavily on white-box optimization and the intrinsic alignment strength of the protected model. To address these limitations, we propose RIB-Guard, a safety-aware information bottleneck defense for black-box LLMs. RIB-Guard learns a token-level masking policy that extracts a minimally safety-sufficient prompt via reinforcement learning using only black-box feedback. In addition, it introduces an independent lightweight safety head to estimate residual jailbreak risk and provide model-agnostic safety guidance during training. The proposed framework jointly balances prompt compactness, benign utility preservation, and residual risk suppression within a unified objective. Experimental results on direct single-turn harmful and benign prompt settings show that RIB-Guard improves jailbreak robustness while maintaining competitive benign utility. By extending information bottleneck-based prompt protection from white-box to black-box settings, RIB-Guard provides a step toward safety-aware information-theoretic front-end defense for black-box LLMs. Full article

Tool-augmented large language model agents are increasingly proposed for network configuration, but routing protocols differ in the control-plane state each commanded router can observe. This difference creates a specific problem for multi-agent orchestration: agents may coordinate more, yet still fail when correct verification depends on peer- or remote-router evidence. We study this interaction through 350 controlled runs on RIP, OSPF, and BGP tasks implemented with FRRouting and Containerlab, comparing a single-agent baseline with multi-agent orchestration patterns across language models. Protocol-centric trace metrics, including spatial coverage, coordination tax, and cross-router verification gap, are combined with intent-property scores and model-balanced bootstrap analysis. The results show that observability explains performance more clearly than orchestration patterns: multi-agent templates trail the baseline on local RIP feedback, show only small and uncertain gains on single-area OSPF troubleshooting, and remain near zero on stricter multi-area OSPF and BGP tasks where peer-side verification gaps are often complete. The main contribution is therefore a protocol-centered account of when agentic orchestration helps, when it adds coordination cost, and why current architectures face a cross-router verification ceiling. Full article

DC–DC converters are ubiquitous in consumer, industrial, commercial, and medical applications. In such voltage-, power-, and area-constrained systems, guaranteeing accurate output voltage remains a key challenge. Investigation of the fundamental cause of steady-state output errors in DC–DC converters, however, is largely absent in the literature. This work identifies systemic voltage offset error as one of the key contributors to steady-state output inaccuracy in PWM and peak–valley-controlled switched-inductor voltage regulators. It uses an insightful reverse-feedback translation framework to quantify the systemic offset as a function of the duty cycle, input voltage, sawtooth amplitude, propagation delays, load conditions, error amplifiers, and comparator. Furthermore, with the derived offset expressions, the paper develops accurate and low-overhead design guidelines to remove systemic errors by aligning the regulator’s steady-state equilibrium with its operating conditions. With the proposed offset “centering” and “elimination” techniques, the systemic error (that accounts for up to 2.1% variation in the steady-state output) is reduced by over 70% when centered and to zero when eliminated at room temperature. Overall, this work provides an insightful and generalized quantification of systemic offsets and describes low-overhead strategies to restore steady-state accuracy in practical PWM, hysteretic and peak/valley-controlled voltage regulators. Full article

Ride-hailing dispatch systems face significant challenges under fluctuating demand and dynamic traffic conditions, where efficient coordination is essential for both platform performance and driver income among large-scale ride-hailing vehicles. This paper constructs a grid-based ride-hailing vehicle dispatch decision model (GRV-DDM), which provides a structured and quantifiable representation of vehicles and orders, effectively capturing spatio-temporal heterogeneity in dynamic traffic environments. Based on this model, a Bi-Level Optimization Multi-Directional Dispatch Decision Algorithm (BO-MDDA) is proposed. At the macro level, evolutionary game theory is employed to adaptively guide collective vehicle strategies toward supply–demand equilibrium, while at the micro level, deep reinforcement learning optimizes individual drivers’ real-time dispatch decisions to maximize long-term profits. A bidirectional feedback mechanism is further designed to integrate macro-level collective intelligence with micro-level individual decision-making. Experimental results across diverse traffic scenarios demonstrate that the proposed approach outperforms classical dispatch algorithms in terms of efficiency and robustness. Full article

This paper presents the design and fabrication of a wideband low-noise amplifier (LNA) covering C-band, using the 0.15 µm GaAs pHEMT process. To achieve both low noise performance and wide matching characteristics, a two-stage cascaded architecture is implemented. In the first stage, circular inductors and an inductive source degeneration technique are employed to minimize the noise figure (NF) while ensuring wideband input matching. Furthermore, an RC feedback structure is incorporated to effectively enhance the stability of the amplifier. The proposed LNA operates under a supply voltage of 3.3 V and a gate bias of 0.35 V, with a total DC power consumption of 69.3 mW. The fabricated MMIC occupies a total chip area of 1.98 mm², including the probing pads. Measurement results demonstrate that the LNA achieves an NF of 0.81–1.09 dB and a gain of over 20.1 dB in the frequency range of 3.3–8.0 GHz. The input and output return losses are maintained over 10 dB and 9.7 dB, respectively. Full article

Protected horticulture moves fragile pots, plug trays, seedlings, harvested products, and carriers through narrow, humid, and crowded spaces. Transport robots must therefore integrate locomotion, perception, localization, handling, placement, scheduling, and human–robot interaction rather than operate as simple carts. This structured narrative review reorganizes evidence from seedling transplanting, nursery operations, harvest support, manipulation, perception, and autonomous navigation around the complete transport chain: target recognition, pickup, loading, loaded navigation, docking, unloading or placement, payload protection, and workflow feedback. The synthesis covers mobile platforms, payload support, perception and localization, motion control, gentle handling, digital support, and fleet coordination. Three barriers remain: short laboratory tests rarely provide season-long evidence; many prototypes are too specialized for variable workflows; and benchmarks seldom combine motion accuracy, handling reliability, payload quality, and resilience. Progress will require modular platforms, robust sensing, payload-safe control, standardized interfaces, and closer co-design between robotics and horticultural operations. Full article

A common hypoxic scenario in tumors involves unresolved acute hypoxia that eventually leads to sustained (chronic) hypoxia. This shift drives a characteristic “HIF switch”, where the key hypoxia-responsive factors change from HIF-1α to HIF-2α over time, and importantly, this switch is closely linked to stemness regulation. However, the mechanisms underlying this switch and its impact on stemness regulation are not yet fully understood. Here, we developed a mechanistic network model integrating the HIF-1/HIF-2 signaling axis with the stemness regulators OCT4 and SOX2. We found the duration and intensity of hypoxia jointly shape the dynamics of HIF-1α and HIF-2α, ultimately regulating OCT4-mediated stemness. Under physioxia, HIF-2α–mTORC2 positive feedback supports the gradual accumulation of HIF-2α toward a modest steady level and low OCT4 expression, corresponding to a primed state. Under prolonged mild hypoxia, the concurrent induction of HIF-1α, albeit at low levels, and accelerated accumulation of HIF-2α elevate OCT4 to intermediate levels, promoting stem-like traits. Under moderate hypoxia, PHD-2-mediated negative feedback triggers pulsatile HIF-1α dynamics, driving a shift toward HIF-2α dominance. Ultimately, cooperative HIF-1α/HIF-2α signaling induces REDD1 and suppresses mTORC1-dependent protein synthesis, pushing OCT4 into a high-expression state associated with differentiation. This work presents a unified framework for understanding how the HIF signaling hierarchy coordinates metabolic and transcriptional programs to direct cell fate across varying hypoxic landscapes. Full article

A new class of modified minimum state error variance (MSEV) state-space based optimal linear quadratic Gaussian (LQG) regulators for closed-loop structures with estimated feedback has been proposed in this article. The negative feedback path is designed as the cascade of the digital LQG regulator and discrete Kalman state observer. The proposed design enables tracking of a time-varying reference input using the predictive control approach. Moreover, the proposed tracking method utilizes a multivariable continuous-time Cauchy state-space model of nonlinear, nonstationary dynamic systems. The resulting control strategy is approximately optimal, as the optimality of the LQG design holds locally for each linearized model around the respective operating point and does not extend to the global nonlinear system. In this sense, starting from the prespecified nominal state trajectory to be tracked, a numerical optimization procedure minimizing the squared tracking error at each step by using the Nelder–Mead direct search simplex algorithm under the required constraints on the input signal has been developed. The LQG regulator and Kalman state observer are designed by utilizing the linear discrete-time state variable models that properly approximate the nonlinear system dynamics across the nominal state trajectory. The performance of the proposed design is validated by simulating a six-degree-of-freedom nonlinear aircraft model across typical flight regimes. Full article

Office openness comprises two physically distinct dimensions—spatial openness and visual openness—yet studies quantifying their independent contributions to cognitive efficiency at the individual level remain scarce. Prior research has predominantly reported group-mean effects, leaving bidirectional individual responses insufficiently examined. This study independently manipulated both dimensions and measured individual-level EEG responses in 24 adults using a 3 × 3 within-subject factorial design. The beta/alpha ratio change rate was computed as an index of cognitive efficiency. Substantial neurophysiological variation across conditions was confirmed in every participant. The absence of significant group-level effects was interpreted not as a lack of environmental influence but as the result of bidirectional individual responses canceling each other out in group averages. Spatial and visual openness induced response ranges of equivalent magnitude at the individual level, and individually optimal conditions were widely distributed across the nine experimental conditions. The correspondence rate between subjective preferences and EEG-identified optimal conditions did not exceed chance, and this bidirectional cancellation mechanism is proposed as an explanation for the contradictory findings that have long characterized open-office research. These results support design strategies that offer diverse combinations of spatial and visual openness within activity-based working environments, paired with feedback systems grounded in objective cognitive performance data. Full article

Fuzzy control decision-making (FCDM) is an intelligent paradigm that leverages linguistic variables to model knowledge, explicitly addressing epistemic uncertainty (incomplete system knowledge) and aleatoric uncertainty (stochastic noise), thereby enabling resolution of multi-objective optimization conflicts in complex industrial systems. Following the PRISMA protocol, this study conducts a systematic literature review of 123 peer-reviewed publications retrieved from IEEE Xplore, Web of Science, ScienceDirect, and Google Scholar over the period 1965–2026, with emphasis on developments in the past 15 years. Existing reviews predominantly focus on isolated subdomains (e.g., scheduling, maintenance, energy systems), lacking a unified cross-scenario synthesis and implementation framework for industrial FCDM. To address scalability challenges such as rule base explosion in high-dimensional spaces, the literature is analyzed with respect to hierarchical fuzzy architectures, rule pruning, and dimensionality reduction techniques. The primary contribution is a structured synthesis of FCDM mechanisms across four industrial domains, combined with a systematic examination of integration with Industrial Internet of Things (IIoT), Digital Twins, and Edge Analytics. Furthermore, a three-stage closed-loop framework is formalized as a unified optimization protocol and modular architecture with technical specifications for Industry 4.0 integration, comprising data preprocessing, fuzzy inference, and optimization-driven decision output with iterative feedback. Comparative evaluation against MILP, MPC, and DRL highlights the conditions under which FCDM provides superior robustness and interpretability. Full article

The interface design of conversational AI mobile applications is shaped by natural language interaction, multi-turn feedback, and dynamically generated content. While these features may reduce certain operational barriers, they can also create new difficulties for older adults in understanding system functions, judging generated results, and recovering from interaction errors. To address these challenges, this study integrates the KANO model and the Analytic Hierarchy Process (AHP) to develop a systematic framework for analyzing interface requirements in conversational AI mobile applications for older users. Field surveys and semi-structured interviews were first conducted to identify 15 interface design requirements. These requirements were then classified through a KANO questionnaire into must-be, one-dimensional, attractive, and indifferent categories, with no reverse requirements identified. On this basis, an AHP hierarchy was established to determine the relative priority of each requirement. The results show that clear functional explanations, interface simplicity, absence of advertising interference, voice interaction, and error-tolerant interaction design are the key factors influencing older adults’ experience with conversational AI interfaces. Basic usability requirements mainly reduce barriers to use, while functional explanations and voice interaction help older users understand system capabilities and task procedures. Error-tolerant interaction further enhances users’ sense of security and control in dynamic and uncertain conversational contexts. These findings suggest that age-friendly design for conversational AI mobile applications should not be limited to isolated adjustments of fonts, icons, or colors. Instead, it should adopt a systematic approach centered on low-complexity interfaces, clear task guidance, interpretable feedback, and recoverable interactions. Based on the classification and weighting results, this study proposes an interface design framework for age-friendly conversational AI mobile applications, providing a reference for requirement analysis, interface optimization, and design decision-making. Full article

Pedestrian trajectory prediction in complex occlusion environments remains a critical challenge for autonomous driving systems. Although high-precision prediction models have achieved notable success, they often entail substantial computational overhead and struggle to maintain both accuracy and physical plausibility under real-world occluded conditions. To address these limitations, this paper proposes OPTP-System, a lightweight prediction framework that integrates YOLOv11 with DeepSORT for robust multi-pedestrian tracking in occluded scenes. An extended Kalman filter (EKF)-based motion prediction module is employed to generate trajectory forecasts, while the EKF-derived prior knowledge guides detection re-searching in occluded regions. Furthermore, feedback from trajectory smoothing refines detection confidence, substantially enhancing the model’s capability for continuous tracking and prediction under severe occlusion. Experimental results under challenging occlusion settings (exceeding 50% occlusion) show that the proposed model reduces ADE and FDE by 30.0% and 29.3%, respectively, compared to state-of-the-art methods. These findings demonstrate that OPTP-System achieves superior prediction accuracy while maintaining computational efficiency, offering a practical solution for reliable pedestrian trajectory prediction in complex traffic environments. Full article

Long noncoding RNAs are involved in diverse biological processes in plants. Our recent study has revealed that an ABA-induced long intergenic noncoding RNA, ARTA, regulates both ABA and drought responses by blocking the nuclear import of a transcription factor, MYB7, through interacting with an importin β-like protein, SAD2. Here, we show that unlike MYB7, ARTA fails to disrupt interactions of SAD2 with the other two R2R3-MYB subgroup 4 members, MYB4 and MYB32. Consequently, the nuclear localizations of MYB4 and MYB32 remain unchanged upon alteration of ARTA expression. Furthermore, ARTA and MYB7 form a self-reinforcing feedback loop during Arabidopsis responses to ABA: ABA treatment induces ARTA expression, which in turn inhibits nuclear accumulation of MYB7, thereby deteriorating MYB7-mediated repression of ARTA and promoting ARTA production. This self-reinforcing feedback regulation elegantly integrates protein relocalization with transcriptional augmentation in the ABA response process, and provides a tunable molecular circuit for plant stress adaptation. Full article

Cognitive radar has been developed to use feedback from its operating environment, obtained from a beam, to make resource allocation decisions by solving optimization problems. Previous works focused on target tracking accuracy by designing an evaluation metric for an optimization problem. However, in practical real-world scenarios, both the target tracking performance of cognitive radar and its operational perspective should be considered. In this study, the usage of an operational risk score in the allocation of radar resources is proposed for a cognitive radar framework. Resource allocation regarding radar dwell time is considered to reflect the operational significance of the target’s priority level. The dwell time allocation problem is solved through Second-Order Cone Programming (SOCP). Numerical simulations are performed to verify the effectiveness of the proposed framework. The results show that the proposed SOCP-based algorithm achieves comparable operational risk estimation performance to conventional methods while using fewer time resources, thereby improving overall resource efficiency in resource-constrained environments. Full article

This work studies a finite-time mean-field type control problem arising from optimal investment under uncertainty with risk management. The problem leads to a nonlinearly coupled system of parabolic equations with temporal and nonlocal interactions. An explicit characterization of the solution to the system is obtained, and a generalized finite difference method (GFDM) combined with an iterative scheme is developed to ensure global temporal consistency of the mean-field feedback during backward computation. Numerical experiments illustrate the accuracy and effectiveness of the proposed approach.In addition, sensitivity studies with respect to the volatility and risk-aversion parameters demonstrate the robustness of the proposed numerical framework under parameter perturbations. Full article