Search Results (219)

Background: Knee osteoarthritis (KOA) is often framed as degenerative knee pain, yet behaves as a decades-long care pathway in which medication, injections, comorbidity, productivity loss, and surgery accumulate into a major economic footprint. This critical narrative review synthesizes pharmacoeconomic and pharmacoepidemiologic evidence across that pathway. Methods: Structured source identification was conducted in PubMed, Web of Science, Scopus, and Google Scholar for publications from 2000 to 2026, with citation tracking. Sources were appraised against predefined critical-interpretation domains and mapped narratively rather than pooled; no meta-analysis was performed. Results: Global Burden of Disease 2019 estimates approximately 364.6 million prevalent KOA cases worldwide. Reported evidence indicates that KOA spending is highly concentrated: in a large U.S. claims analysis, knee arthroplasty was performed in approximately 8.8% of patients yet accounted for 61.5% of KOA-related costs, whereas hyaluronic acid represented 3.0% of overall costs; the remaining pathway burden was distributed across years of outpatient care, analgesics, injections, and other nonsurgical utilization. Medication and injection findings were stage- and phenotype-dependent, and observational studies associated opioid exposure with higher fall risk, healthcare utilization, and cost. Intra-articular hyaluronic acid was repeatedly associated with longer time to arthroplasty, interpreted here as an association limited by confounding and immortal-time bias, not a causal effect; platelet-rich plasma value remained price- and durability-sensitive. Conclusions: KOA economics resembles an iceberg—arthroplasty is the visible peak, while the submerged mass is years of pathway-level care. Value-based policy should measure the full pathway, not the surgical episode, using linked claims, registries, patient-reported outcomes, and productivity data. Full article

Artificial intelligence (AI) agents increasingly need credentials such as application programming interface (API) keys and Secure Shell (SSH) credentials, but placing those secrets in the agent process exposes them to prompt injection, tool misuse, and exfiltration through ordinary agent outputs. We present CapSeal, a capability-based broker that replaces direct secret access with session-bound, non-exportable handles. Agents request policy-evaluated actions, while the broker performs credential-bearing Hypertext Transfer Protocol (HTTP) and SSH execution through typed executors with schema validation, replay protection, revocation epochs, and tamper-evident audit logging. We extend this design to hosted settings where the broker operator is not trusted with tenant secrets. Our main contribution is operator-blind secret mediation: a split-broker architecture in which a small trusted tenant gateway cooperates with an untrusted operator service that stores the master secret only as a fully homomorphic encryption (FHE) ciphertext and evaluates per-request derivations without decrypting it. We formalize the model and prove computational operator blindness from indistinguishability under chosen-plaintext attack (IND-CPA) security of the FHE scheme, together with conditional capability binding for any secure pseudorandom function/message authentication code (PRF/MAC) instantiation. We implement an end-to-end TFHE-rs prototype that exercises split-broker derivation, multi-tenant revocation and rate limiting, audit integration, and HTTP/SSH mediation. The prototype uses a non-cryptographic homomorphic stand-in and measures the cost of crossing the operator-untrusted boundary at about 9 s per request, roughly 17 million times slower than the plaintext path. We also give LowMC and Rasta transciphering designs and compare FHE with trusted execution environment (TEE)- and secure multiparty computation (MPC)-based alternatives, positioning each trust boundary by assurance and performance. Full article

Improving regional economic resilience is a point addressed in the sustainable development goals (SDGs; i.e., SDG 8 and SDG 11). The Yangtze River Delta (YRD) has demonstrated excellent economic resilience during the COVID-19 pandemic, largely due to the persistent green transition of the YRD in the past two decades. This paper uses a single-case method combined with the perspective of evolutionary economic geography to systematically investigate the process of green transition in the YRD (2000–2023) at both vertical and horizontal levels and proposes an integrated multi-dimensional system framework to reveal the collaborative logic of the overall green transition action and the internal mechanism of enhancing economic resilience in the YRD. The findings indicate that the combination of external factors such as contradiction change, magnifying crises, economic stabilization, and policy steering has driven the historical inevitability of green transition in China. Under such conditions, the YRD not only completed development in terms of primitive accumulation of space (coordinated development, i.e., chassis), industry (orderly upgrade, i.e., engine), and governance (equal supply, i.e., lubricant) earlier but also ensured the stability of this triangle, injecting sustained strong momentum into the rapid recovery of the economy under the impact. The solidification of green concepts further enhances the sustainability and strength of the YRD’s economic resilience. These findings provide beneficial experience on how to resume production after the pandemic or lay out cities in developing countries that are still in rapid urbanization in advance. Full article

Water scarcity in Chile has been exacerbated by a decline in precipitation and an increase in water demand. This has prompted a search for strategies to increase water supply, whether through aquifer recharge or reservoir construction. In this study, aquifer recharge was evaluated through rainwater harvesting systems (RWHS) and direct injection into rural wells in the Ñuble Region. Three wells were selected in the Ñuble Region (Ñiquén, San Carlos, and Coihueco) using hydrogeological and operational criteria. To characterize the hydrogeology of the area, local piezometric data, geophysical surveys using electrical resistivity tomography (ERT), and seismoelectric tests were considered. This enabled the identification of aquifers with water levels between 2.6 and 23 m depth across the different geological units of the territory. The hydrological design was based on a frequency analysis of annual precipitation (1991–2020), which yielded design rainfall values between 442 and 694 mm. The implemented RWHS demonstrated injection capacities between 0.9 and 1.4 L·s⁻¹. The results show that rainwater harvesting combined with direct aquifer recharge represents a viable alternative for improving water security, with potential for territorial scaling through regional public policies. Full article

Accurate carbon price forecasting is essential for emission trading risk management and low-carbon investment decisions. In existing decomposition-prediction frameworks, secondary decomposition targets are typically selected based on statistical complexity rather than domain-informed causality, and standard Long Short-Term Memory (LSTM)-Transformer architectures discard the cell state that encodes long-term temporal memory. These limitations are particularly pronounced where energy-driven causal structures and regime-switching volatility coexist. This study proposes Causal State-driven Network (CaStNet), an intelligent forecasting framework with two core innovations. A Policy-Causality-guided Residual Secondary Decomposition (PCRSD) module replaces entropy-based criteria with Granger causality to select intrinsic mode functions (IMFs) exhibiting significant energy-carbon causal linkages for targeted variational mode decomposition (VMD). A Cell-State-Driven Dual-function Attention (CSDA) mechanism repurposes the LSTM cell state for simultaneously injecting long-term memory into the Transformer and employing the cell-state differential velocity as a volatility proxy to adaptively regulate Top-k attention sparsity. The Artificial Lemming Algorithm (ALA) globally co-optimizes decomposition dimensions and attention boundaries. A Shapley Additive exPlanations (SHAP)–Local Interpretable Model-agnostic Explanations (LIME) interpretability analysis reveals horizon-dependent driver transitions from short-term autoregressive momentum to long-term energy fundamentals, uncovering threshold nonlinearities in energy-carbon transmission channels. Validation on the Shanghai market (2013–2025) achieves point-forecast RMSE = 0.8326 and R² = 0.9777, outperforming all twelve benchmark models. Cross-market testing on the Hubei market yields R² = 0.9487, and expanding-window five-fold cross-validation on the Shanghai dataset yields mean R² = 0.9704, jointly confirming generalization robustness. Full article

Tool-calling LLM agents are vulnerable to indirect prompt injection: externally retrieved data can redirect tool calls without system-prompt access, and prompt-level defences leave three harm classes undefended (path traversal, user-guided exfiltration, high-frequency tool abuse). We present a Policy Enforcement Point (PEP) that intercepts at the tool-call boundary with declarative rules over a cross-step information-flow label system (source integrity, data sensitivity) and a synchronous SHA-256 hash-chained audit log. On a controlled dataset across four attack classes, the full system cuts the attack success rate (ASR) from 40.0% to 5.0% (deepseek-v4-pro, five repeats) versus 35.0% for the strongest prompt-only baseline; disabling cross-step label propagation raises the call-level false-negative rate by 26.4 points. The 30.0% task-level false-positive rate is dominated by by-design least-privilege capability-token denials, not rule false positives—an expanded 30-task benign set yields 0/30 rule false positives under scripted isolation. A conservative-DS mitigation (intent-taint) closes the constructed denied-read reconstruction blind-spot variant (ASR 100% to 0%) at no cost on standard workflows. The audit log detects all three tested tamper classes; the in-process enforcement overhead is sub-millisecond per call. Across four further backends, ASR drops under the full system, though LLaMA-3.3-70B retains 16.7% (a rule-coverage gap). A preliminary run over a real MCP stdio transport (an official filesystem server) shows the mechanism operates at a real boundary with a sub-millisecond execution-path increment. We frame these as mechanism-coverage evidence on a controlled benchmark, not a deployability claim for production MCP workloads. Code, data, and metrics are openly available in the replication repository. Full article

Fault-injection campaigns are widely used to evaluate silent data corruption (SDC) in AI hardware, but exhaustive campaigns over workloads, dataflows, processing elements, and datapath roles are expensive. This paper presents an architecture-level risk-guided fault-injection prioritization method for systolic AI accelerators. The method ranks candidate transient functional perturbations before downstream validation, with the goal of enriching the discovery of candidates that produce a thresholded relative-output-error outcome under a limited validation budget. The evaluation uses a fixed candidate fault pool: all ranking policies score the same 21,000 candidate faults across 30 workload/dataflow/array configurations, corresponding to five GEMM-derived workloads, three array sizes, and two dataflows. Fault magnitudes are sampled once per candidate and are independent of all ranking scores. Candidate faults are modeled as transient architecture-level perturbations in MAC, accumulator, or forwarding paths. The proposed full-risk score combines activity, composite spatial stress, tensor sensitivity, and a path-class weight. In the proposed architecture-level simulation environment and under the fixed-pool protocol, the proposed method achieves the highest mean top-10% SDC-proxy lift, AUPRC, NDCG@10%, and rank correlation with relative output error among the evaluated principle-based ranking policies. At the calibrated threshold, it achieves a mean top-10% lift of 5.65× [4.91, 6.38], compared with 4.61× for AVF-like exposure and 4.33× for output sensitivity. Paired configuration-level tests, threshold sensitivity, and outcome-model sensitivity analyses characterize the result while showing that the proposed score is not universally dominant under every synthetic outcome assumption. The method is intended as a front-end architecture-level screening tool for validation prioritization, not as a replacement for RTL, gate-level, FPGA, or silicon reliability signoff. Full article

Modern power grids with high renewable energy penetration are vulnerable to fast voltage disturbances caused by grid faults. Among these, voltage sags are critical because they develop within milliseconds and require rapid reactive current support to maintain grid stability and power reliability. Reinforcement learning has previously shown potential for reactive current injection control during voltage sag events due to its fast response and adaptability to changing system conditions. However, existing approaches rely on separate policies for specific subsets of the operating space, which limits their ability to provide optimal actions when the system operates across broader or combined state regions. To address this limitation, this paper proposes a unified Soft Actor–Critic (SAC) target policy trained over the full state and action space by integrating multi-source transfer learning with potential-based reward shaping approach. Results show that the proposed multi-source transfer approach enables the target agent to converge faster and reach a higher reward solution than the baseline SAC and single-source transfer approach. The trained policy also improved prediction accuracy, achieving reactive-current errors below 0.2 A with respect to the ground-truth reference generated through extensive simulations over the full observation and action space. The reference follows the grid-code requirement for minimum reactive current injection during faults and provides a benchmark for evaluating prediction accuracy. This can help distributed generation sources respond more effectively during severe perturbations such as voltage sags, support voltage recovery, and reduce the risk of cascaded disconnections that could lead to unwanted blackouts. Additionally, the inference execution time is also sufficiently fast to satisfy the response-time requirement of voltage sag events, confirming the real-time feasibility of the proposed controller. Full article

Background: Infective endocarditis (IE) requires 4–6 weeks of intravenous antimicrobial therapy, and timely transition to outpatient parenteral antimicrobial therapy (OPAT) allows clinically stable patients to complete treatment outside the hospital. Because OPAT requires home infusion services or post-acute facility placement that typically depend on coverage, insurance status may strongly influence length of stay (LOS); national data on this association in IE remain limited. Methods: We performed a retrospective cross-sectional analysis of the 2016–2019 National Inpatient Sample (NIS) using ICD-10-CM codes I33 and I38 to identify adult IE hospitalizations. Patients were classified as insured (Medicare, Medicaid, or private insurance) or uninsured (self-pay or no charge). Outcomes included mean and prolonged LOS (>14 and >28 days), in-hospital mortality, discharge against medical advice (AMA), and hospitalization costs. Comparisons used chi-square and Student’s t-tests with appropriate NIS survey weighting. Multivariable Gamma regression (LOS, cost) and logistic regression (binary outcomes) were performed, adjusting for age, sex, race/ethnicity, income quartile, injection drug use (IDU), Elixhauser Comorbidity Index, and hospital characteristics, with an insurance × IDU interaction term. Results: Of 87,211 weighted IE hospitalizations, 81,667 (93.6%) were insured and 5544 (6.4%) were uninsured. Uninsured patients were younger (mean age 40.1 vs. 59.4 years) with lower comorbidity burden but higher injection drug use (IDU) prevalence (38.7% vs. 15.5%). Mean LOS was longer among the uninsured (15.5 vs. 12.4 days, p < 0.001); LOS > 14 days occurred in 35.8% vs. 26.6%, and LOS > 28 days in 18.5% vs. 9.2% (both p < 0.001). AMA discharge was four-fold higher among the uninsured (22.2% vs. 5.5%, p < 0.001), while unadjusted in-hospital mortality was similar (9.0% vs. 9.4%, p = 0.32). LOS and AMA disparities persisted in both IDU and non-IDU subgroups, with a six-fold AMA disparity among non-IDU patients (15.2% vs. 2.5%). Based on multivariable analysis, uninsured status remained independently associated with prolonged LOS > 28 days (adjusted odds ratio [aOR] 1.46, 95% CI 1.30–1.65), AMA discharge (aOR 3.51, 95% CI 3.10–3.97), and—after accounting for age and comorbidity differences—higher in-hospital mortality (aOR 1.25, 95% CI 1.10–1.43). Conclusions: Uninsured adults hospitalized with IE experienced longer stays, markedly higher AMA rates, and—after adjustment for age and comorbidity—higher in-hospital mortality than insured patients. These findings are consistent with nonclinical barriers to discharge—particularly limited OPAT and post-acute care access—and suggest that the younger, less comorbid profile of uninsured patients masks an underlying outcome disparity. The results identify uninsured IE patients as a population that may benefit from alternative care models and policy reforms expanding safe post-acute antimicrobial therapy. Full article

Secure data collaboration among mutually distrustful organizations requires more than encrypted storage: it also needs accountable ownership control, auditable access governance, privacy-preserving transaction execution, and reliable settlement when data are exchanged as digital assets. This paper proposes TrustTrade, a unified multi-party secure data management and transaction framework designed for cross-organization data sharing, trading, and compliance-sensitive analytics. TrustTrade integrates policy-bound data capsules, a tamper-evident provenance ledger, adaptive threshold escrow, verifiable data-payment settlement, and selective audit with revocation rebinding. On four real-dataset-derived workloads, TrustTrade reaches a $90.4$–$94.8\%$ settlement rate, with a $92.5\%$ average that is $6.4$ percentage points higher than the strongest baseline average. Under adversarial request injection, TrustTrade reduces unauthorized release to $0.31\%$ and atomicity violation to $0.38\%$, corresponding to $93.6\%$ and $93.0\%$ reductions compared with Plain-Market, respectively; compared with Fixed-Escrow, unauthorized release is reduced by $77.4\%$. TrustTrade also achieves $96.7\%$ dispute-resolution accuracy while maintaining practical settlement latency. These results indicate that jointly designing secure data management and secure data transaction protocols offers a practical path toward trustworthy multi-party data ecosystems. Full article

This paper proposes a reinforcement learning-assisted active disturbance rejection control (ADRC) framework for a nonlinear quadrotor unmanned aerial vehicle (UAV). Conventional ADRC controllers are designed for the quadrotor altitude and attitude channels. To evaluate robustness under disturbance-intensive conditions, a composite external disturbance is injected into the roll-channel dynamics. A Deep Deterministic Policy Gradient (DDPG)-based adaptive tuning mechanism is integrated into the roll-channel ADRC for the nonlinear state error feedback (NLSEF) gain adaptation, while fixed-parameter ADRC is retained for the remaining three channels. Without requiring system linearization and prior knowledge of disturbance models, the reinforcement learning agent learns an optimal gain adaptation policy directly through interaction with the nonlinear roll subsystem. Quantitative simulations demonstrate superior roll-axis disturbance rejection, leading to 90% faster settling time, the root mean square (RMS) control effort being reduced by 5.1%, and a 7.6% peak input suppression compared to conventional ADRC. The learning-based adaptation maintains comparable tracking accuracy across all channels while significantly improving transient recovery and control smoothness in the most disturbance-sensitive axis, validating selective reinforcement learning integration for robust nonlinear quadrotor flight control. Full article

Using panel data from Chinese cities spanning 2010–2023 and leveraging the natural experiment provided by the establishment of the National Big Data Comprehensive Pilot Zone (NBDPZ), we employed the difference-in-differences (DID) method alongside double machine learning (DML) to systematically examine how these policies influence urban economic resilience efficiency. The empirical results demonstrate that the NBDPZ significantly enhances urban economic resilience efficiency. This finding is robust under parallel trend and placebo tests, confirming that the improvement is a policy-driven causal effect. Mechanism analysis reveals that the policy enhances urban economic resilience efficiency primarily by promoting the upgrading and rationalization of industrial structure to consolidate the micro-foundation of sustainable economic transformation; increasing innovation output to facilitate the sustainable accumulation of knowledge capital; and enhancing urban entrepreneurial activity to inject sustainable endogenous vitality into the economic system. Heterogeneity analysis indicates that the positive effects are more pronounced in eastern and western regions, second-tier cities, and cities with lower industrial agglomeration, better digital infrastructure, and stronger legal and regulatory environments. The study’s findings offer both theoretical support and practical guidance for refining the policy framework of the NBDPZ policy and promoting sustainable urban economic development. Full article

Large language models (LLMs) offer a flexible interface for human–machine interaction, but their direct use in embedded control remains difficult because low-cost microcontrollers cannot host such models locally and unconstrained language generation is not physically grounded. This paper presents EdgeTalk-MCU, a local host–microcontroller framework for low-latency natural-language control of resource-constrained devices. The system couples a locally deployed LLM on the host side with an ESP32-S3 microcontroller through a lightweight serial protocol and closes the loop with real-time state feedback. The reported end-to-end decision latency of ∼0.15 s refers to the host-side inference pipeline; physical platform latency additionally includes UART round-trip and servo actuation overhead. The design combines two complementary mechanisms: a state-aware prompt constraint that injects task progress and physical state into the host-side policy, and a runtime shield that enforces hard execution consistency before actuation. This decomposition separates raw policy quality from executed safety. Across representative obstacle scenarios in simulation, unshielded controllers remain unreliable—LLM-only and Prompt-only exhibit collision rates of 30.6% and 26.5%, respectively, in the Sudden Obstacle setting—whereas both shielded methods reduce collision to 0%. An ablation study confirms that the runtime shield is the decisive safety mechanism; the state-aware prompt constraint contributes primarily at the raw-proposal level by reducing the fraction of unsafe proposals submitted to the shield, rather than by independently guaranteeing safe execution. Hardware-in-the-loop (HIL) validation on a physical ESP32-S3 platform confirms that the same qualitative pattern holds under real sensing and communication conditions. Full article

This paper proposes a lightweight and adaptive anomaly detection framework for Internet of Things (IoT) time-series data that progressively combines statistical filtering with reinforcement learning (RL)-based decision mechanisms. Three classical statistical filters, Hampel, interquartile range (IQR), and Z-score, are initially evaluated under controlled IoT anomaly scenarios. While fixed-parameter configurations perform well under specific conditions, their performance degrades in non-stationary and heterogeneous environments. To address this limitation, a tabular Q-learning agent is introduced to dynamically select both filtering methods and their associated parameters according to scenario-specific signal characteristics. By extending the action space to include joint filter and parameter selection, the framework improves adaptability while reducing the need for manual tuning. A multi-agent reinforcement learning (MARL) formulation is further introduced to support distributed learning across heterogeneous IoT environments. The framework is additionally evaluated using real-world IoT temperature data augmented with controlled anomaly injection, enabling reproducible benchmarking under partially realistic sensing conditions. Experimental results show that both RL and MARL maintain stable detection performance across heterogeneous sensor streams. While MARL does not systematically outperform the single-agent approach in detection accuracy, it improves scalability and supports scenario-specific policy specialization, which is particularly relevant for distributed IoT deployments. Overall, the proposed approach provides a lightweight, interpretable, and computationally efficient solution for adaptive anomaly detection in resource-constrained IoT systems. Full article

Evaluating intrusion detection methods at the level of individual MITRE Adversarial Tactics, Techniques, and Common Knowledge (ATT&CK) for Industrial Control System techniques requires Operational Technology traffic in which each attack sequence carries its MITRE technique identifier as ground truth. Publicly available Industrial Control System datasets either provide coarse attack-versus-benign labels (SWaT, WADI, CIC-APT-IIoT) or require ex-post technique reconstruction from CALDERA operation logs, and therefore do not support per-technique benchmarking. We describe one primary contribution and two supporting contributions, demonstrated on one Modbus/Raspberry-Pi programmable logic controller/CALDERA/convolutional bidirectional Long Short-Term Memory autoencoder (CNN-BiLSTM-AE) use case. The primary contribution is an in-orchestrator labelling methodology for per-technique-labelled Industrial Control System attack capture. Its single load-bearing property is that the campaign orchestrator owns the label primitive and writes each per-sequence technique identifier into the capture artefact at injection time, eliminating ex-post log-to-packet alignment. The first supporting contribution is a protocol-aware detection pipeline. Its load-bearing architectural choice is a priority-ordered protocol router that dispatches each labelled flow to a per-protocol detector plug-in (protocol-aware features here, with generic-flow features admissible as an alternative plug-in policy on the same router). The second supporting contribution is a suite of four reproducible CALDERA chains (three Information-Technology-to-Operational-Technology kill chains plus one enterprise-side control) that exercise the labelling methodology end-to-end and the detection pipeline along complementary detection paths. All three contributions are platform-independent: any ATT&CK-aligned emulator and any fieldbus protocol can host the labelling methodology, and any detector trained on an admissible feature space can plug into the router. The dataset contains 40,000 benign and 9997 attack Modbus sequences spanning four ATT&CK techniques (T0802 Automated Collection, T0831 Manipulation of Control, T0836 Modify Parameter, T0846 Remote System Discovery). On this dataset, the CNN-BiLSTM-AE reaches a 100% true-positive rate (TPR) at the 98th-percentile benign threshold across all four techniques and a 99.7% overall TPR at the tighter 99.5th-percentile threshold, with per-technique TPR between 96.1% (T0836 Modify Parameter) and 100% (T0802 Automated Collection, T0846 Remote System Discovery). Across the four CALDERA chains, the Modbus autoencoder produces 234 protocol-layer detections and the Security Information and Event Management (SIEM) rule set produces 30 alerts, with per-chain tactic coverage between 0.714 and 0.786 and CALDERA-ability success rates between 0.800 and 0.857. Full article