A Multi-Agent Reinforcement Learning Framework for Public Health Decision Analysis.

Studies have documented high human immunodeficiency virus (HIV) prevalence among transwomen in the United States; however, to our knowledge, no studies have documented trends in HIV prevalence in this population. We used respondent-driven sampling to sample transwomen in San Francisco for 3 HIV prevalence and behavioral surveys in 2010, 2013, and 2016. Our analysis of point estimates and trends were weighted for the sampling method. Human immunodeficiency virus prevalence by serological testing in the survey was 38.8% (95% confidence interval [CI], 32.4-45.2), 33.7% (95% CI, 25.9-41.5), and 31.6% (95% CI, 12.2-38.1) in 2010, 2013, and 2016, respectively. Disparities in higher HIV prevalence by black, Latino, and Asian race/ethnicity and lower education level persisted through 2016. Based on a statistical test for trend, HIV prevalence among transwomen has remained high and stable from 2010 to 2016. Human immunodeficiency virus infection is still highest at 31.6% compared to any other group in San Francisco. We also observed that older transwomen had significantly higher odds of living with HIV than younger women over the last 2 waves of data collection. Taken together, these trends suggest that there is declining incidence of new HIV infections among low-income transwomen in San Francisco. Moreover, among transwomen, HIV disproportionately affects transwomen of color.

The burden of human immunodeficiency virus (HIV) is high among gay, bisexual, and other men who have sex with men (MSM). High HIV prevalence, lack of awareness of HIV-positive status, unprotected anal sex, and increased viral load among HIV-positive MSM not on antiretroviral treatment contribute substantially to new infections among this population. CDC analyzed data from the National HIV Surveillance System (NHSS) to estimate the percentage of HIV diagnoses among MSM by area of residence and data from the National HIV Behavioral Surveillance System (NHBS) to estimate unprotected anal sex in the past 12 months among MSM in 2005, 2008, and 2011; unprotected discordant anal sex at last sex (i.e., with a partner of opposite or unknown HIV status) in 2008 and 2011; and HIV testing history and the percentage HIV-positive but unaware of their HIV status by the time since their last HIV test in 2011. This report describes the results of these analyses. In all but two states, the majority of new HIV diagnoses were among MSM in 2011. Unprotected anal sex at least once in the past 12 months increased from 48% in 2005 to 57% in 2011 (p<0.001). The percentage engaging in unprotected discordant anal sex was 13% in 2008 and 2011. In 2011, 33% of HIV-positive but unaware MSM reported unprotected discordant anal sex. Among MSM with negative or unknown HIV status, 67% had an HIV test in the past 12 months. Among those tested recently, the percentage HIV-positive but unaware of their infection was 4%, 5%, and 7% among those tested in the past ≤3, 4-6, and 7-12 months, respectively. Expanded efforts are needed to reduce HIV risk behaviors and to promote at least annual HIV testing among MSM.

Smart transportation networks are challenged by a critical challenge of traffic congestion and inevitably requires a complex package of management solutions. This paper introduces a Secure Multi-Agent Reinforcement Learning (MARL) framework that optimizes traffic whilst protecting from malicious threats. Traffic Prediction Dataset accessible through Kaggle is used as the foundation for training reinforcement learning agents from four junctions who cooperate to manage traffic signals. The widespread preprocessing procedures that also dealt with data gaps in addition to dimension transformation and value normalization generated genuine traffic prediction models. The Multi-Agent Reinforcement Learning framework reduced the mean waiting time by 45.7% with a 47.0% traffic passing rate higher than that achieved using standard fixed-duration signal control techniques. Homomorphic encryption, when integrated with blockchain protocols, instituted novel security protocols for agent to agent communication, and this yielded a reduction of adversarial attacks below 90% levels. The implementation framework provided a 39.3% reduction in the fuel consumption when implemented in real-time urban traffic management systems. Secure Multi-Agent Reinforcement Learning (MARL) systems show their potential aptitude to propel smart transportation networks in the while providing increased efficiency and security as well as sustainability effectiveness.

Financial institutions are increasingly dealing with difficulties in preserving transaction integrity due to escalating digital fraud and extensive cloud-based operations. Conventional machine learning and rule-based detection systems frequently have latency in replies, restricted adaptability, and an absence of verifiable audit trails. The proposed study presents a Hybrid Multi-Agent Reinforcement Learning (MARL) and Blockchain framework to overcome existing limitations, integrating adaptive multi-agent coordination for real-time fraud detection with a permissioned blockchain ledger to guarantee secure, tamper-proof validation of agent decisions. MARL agents adaptively acquire optimal detection strategies through incentive shaping and inter-agent communication, while blockchain consensus and smart contracts substantiate decision proofs for regulatory adherence. The experimental evaluation of the Kaggle Financial Transactions dataset demonstrates improved accuracy relative to baseline models: Accuracy 0.989, Precision 0.981, Recall 0.987, F1-Score 0.984, AUC-ROC 0.992, with negligible latency (1.8 ms/transaction). The comparative research reveals a persistent superiority over Random Forest, XGBoost, CNN- LSTM, and Transformer-based classifiers, showcasing robustness, scalability, and real-time application. The suggested study enhances the area by merging reinforcement learning adaptability with blockchain security, establishing a realistic, high-throughput, and reliable framework for cloud- based financial systems, while presenting a model for future research in agentic AI-driven fraud prevention.

Black women are disproportionally affected by human immunodeficiency virus (HIV). This study investigates factors associated with newly identified HIV infection among previously self-reported HIV negative or unknown status black women living in high risk areas (HRAs) of New York City (NYC). Heterosexuals residing in or socially connected to NYC HRAs were recruited using respondent driven sampling for participation in the United States Centers for Disease Control-sponsored National HIV Behavioral Surveillance System in 2010. Eligible individuals were interviewed and offered an HIV test. The analysis reported here focused on black women with valid HIV results who did not report being HIV positive, and examined factors related to HIV infection in this group. Of 153 black women who did not report being HIV positive at enrollment, 15 (9.8%) tested HIV positive. Age ≥40 years, ever injected drugs, and in the last 12 months had unprotected vaginal sex, exchange sex, last sex partner used crack, non-injection crack use, and non-injection heroin use were significantly associated with HIV infection (p<0.05). Only ever injected drugs (prevalence ratio: 5.1; 95% confidence interval: 2.0, 12.9) was retained in the final model. Black women who had reported being either HIV negative or unaware of their serostatus had high HIV prevalence. Efforts to identify and treat HIV positive black women in HRAs should target those with a history of injection drug use. Frequent testing for HIV should be promoted in HRAs.

Efficient parking distribution is crucial for urban traffic management; nevertheless, variable demand and spatial disparities raise considerable obstacles. Current research emphasizes local optimization but neglects the fundamental challenges of real-time parking allocation, resulting in inefficiencies within intricate metropolitan settings. This research delineates two key issues: (1) A dynamic imbalance between supply and demand, characterized by considerable fluctuations in parking demand over time and across different locations, rendering static allocation solutions inefficient; (2) spatial resource optimization, aimed at maximizing the efficiency of limited parking spots to improve overall system performance and user satisfaction. We present a Multi-Agent Reinforcement Learning (MARL) framework that incorporates adaptive optimization and intelligent collaboration for dynamic parking allocation to tackle these difficulties. A reinforcement learning-driven temporal decision mechanism modifies parking assignments according to real-time data, whilst a Graph Neural Network (GNN)-based spatial model elucidates inter-parking relationships to enhance allocation efficiency. Experiments utilizing actual parking data from Melbourne illustrate that Multi-Agent Reinforcement Learning (MARL) substantially surpasses conventional methods (FIFO, SIRO) in managing demand variability and optimizing resource distribution. A thorough quantitative investigation confirms the strength and flexibility of the suggested method in various urban contexts.

Underwater monitoring is extremely complex due to the lack of a global localization system, limited communication and environmental factors such as turbidity and darkness that limit visibility, affecting control and situational awareness. Typically, monitoring relies on a single autonomous underwater vehicle (AUV) or a set of independent AUVs; techniques which are prone to failure as they rely only on onboard odometry and sensors, making missions vulnerable to malfunctions, damage, and noise. To address these challenges, we propose a Multi-Agent Reinforcement Learning (MARL) framework to enable cooperation among multiple AUVs, mitigating the limitations of the underwater environment. Our in-silico solution focuses on a group of robots learning a strategy to follow a partially hidden underwater pipe without global localization, while dealing with environmental disturbances affecting sensors and actuators. The numerosity of the agents, and most importantly their collaboration, helps overcome underwater visibility constraints. By sharing relative position information of neighboring agents with respect to the pipe, navigation is improved. By introducing quantitative measures for pipe exploration, we show that cooperation significantly enhances system performance compared to independent agents. Emerging collaboration among robots allows the swarm to complete pipe inspections faster and more efficiently than non-cooperative baseline models of non-interacting agents, even under extremely reduced visibility scenarios. Moreover, single agents also benefit from cooperation, learning effective policies more quickly and covering a longer portion of the pipe. Finally, our model guarantees explainability. We analyze learned strategies and provide a visualization method that allows the interpretation of the learned policies. • Reinforcement learning is applied to pipeline following by underwater robotic agent. • In conditions of poor visibility, a single agent is not able to complete the mission. • In contrast, multi-agent team completes this task using reinforcement learning. • Swarm collaboration enables faster and more efficient task completion. • Collaboration enables agents develop more efficient individual strategies.

To guarantee heterogeneous performance requirements of diverse vehicular services, it is necessary to design a full cooperative policy for both vehicle to infrastructure (V2I) links and vehicle to vehicle (V2V) links. This paper investigates how to improve the quality of experience (QoE) of the V2I users for video services while satisfying the delay requirements of both V2I and V2V links. In specific, a QoE maximization problem is formulated with consideration of vehicular collaboration where task offloading decision, channel reuse decision and power allocation of V2V users are all included. A multi-agent reinforcement learning (MARL) framework is then designed, where a new reward function is proposed to evaluate the utility of the considered network. Thereafter, a proximal policy optimization approach is proposed to enable each V2V user to learn policy individually with the shared global network reward. The effectiveness of the proposed approach is finally validated with comparison of other baseline approaches through extensive simulation experiments.

As a significant issue in the development of Smart City, traffic light control is closely related to public transport. Multi-agent reinforcement learning is a feasible paradigm to control decentralized systems. In this work, we design a recurrent communication module for traffic light control systems. Aiming at lightweight decentralized learning, we decouple the interaction among entities, using an improved recurrent neural network to share messages. Regarding traffic lights as a partially observable Markov Decision Process, we develop a complete reinforcement learning model. Every light receives messages from its neighbors through the improved recurrent neural network. For the further lights, its neighbor lights can play the role of intermediation to transmit messages. By our communication framework, traffic lights can learn their control policies in a decentralized way, and distribution control can be executed through a communication channel. Several simulated experiments show that our method outperforms the fixed time method and the existing reinforcement learning-based frameworks under different traffic demands.

Multi-agent reinforcement learning for edge information sharing in vehicular networks

Aims: Provision of constant and safe blood has been a public health challenge in Sub-Saharan Africa with a high prevalence of transfusion-transmissible infections (TTIs). This study aimed at determining the seroprevalence of the Human Immunodeficiency Virus (HIV) among prospective blood donors at two Hospitals (government and private-owned) in Rivers State, and also to relate some demographic studies to the screening results.&#x0D; Study Design: Cross-sectional study.&#x0D; Place and Duration of Study: Two Hospitals (a government-owned and private-owned) in Rivers State, Nigeria, between January 2018 and April 2019.&#x0D; Methodology: Two hundred and eighty-two (185 males and 97 females) blood donors were recruited for this study. Sera samples were screened for antibodies to HIV-1 and -2 using enzyme-linked immunosorbent assay (ELISA) based kits following the manufacturer’s description.&#x0D; Results: Of the 282 screened prospective donors (males and females) in this study, the overall prevalence of HIV from both hospitals was 6.0% with a seronegativity of 94.0%. There was a significant relationship (p &lt;0.05) between the overall seroprevalence of HIV concerning gender (p-0.006) all other demographics had no significant association with HIV. Age group 21-30 had the highest prevalence of HIV (53.80%). Donors with tertiary education had the highest prevalence rate (52.90%) of HIV. About marital status, the unmarried donors had higher HIV prevalence (64.70%) when compared with the married donors (35.30%). However, family donors had the highest prevalence of HIV (52.90%). Finally, concerning occupation, students had a higher HIV prevalence (47.10%).&#x0D; Conclusion: The seroprevalence of HIV in Port Harcourt, Rivers State, Nigeria was high. This shows that HIV remains a threat to safe blood transfusion and public health in Nigeria. Strict adherence to selection criteria and algorithm of donor screening is hereby advocated.

Human immunodeficiency virus (HIV) infection continues to expand worldwide, and a significant proportion of infection is still undiagnosed. Recent studies have addressed the impact and feasibility of 'opt-out' HIV screening in Emergency Departments (EDs) in urban settings at high HIV prevalence, whereas little is known about the yield of implementing 'targeted' HIV testing, especially in low-prevalence areas. The present study undertakes a scoping review of research carried out on the implementation of targeted HIV screening of adult in EDs to determine the impact, feasibility and acceptability of HIV testing in different HIV prevalence settings. Online databases (EMBASE, MEDLINE) were used to identify papers published between 2000 to 2020. A three-concept search was employed with HIV (HIV, Human immunodeficiency virus infection, HIV infections), targeted testing (Target, screening or testing) and emergency medicine (Emergency Service, emergency ward, A&E, accident and emergency or Emergency Department) (28th February 2020). Only full-text articles written in English, French, Spanish or Italian and using impact and/or feasibility and/or acceptability of the program as primary or secondary outcomes were analysed. The search provided 416 articles. Of these, 12 met inclusion criteria and were included in the final review. Most of the included studies were carried out in the United States (n=8; 67%) and in areas of high HIV prevalence (n=11; 92%). Three (20%) were randomized control studies. While the rate of newly diagnosed HIV cases varied widely (0.03-2.2%), likely due to methodological heterogeneity between studies, the linkage of new HIV diagnosis was often high (80-100%) and median CD4+ cell count was always greater than 200 cells per microliter. Targeted HIV screening was found to be cost-effective (out of 2 studies) and well accepted by participants (out 2 studies). Targeted HIV screening at the ED can be impactful, feasible and well accepted, but often requires extra funding and staff. Most previous work has focused on areas of high disease prevalence.

BackgroundThere is still a paucity of evidence on the outcomes of coronavirus disease 2019 (COVID-19) among people living with human immunodeficiency virus (PWH) and those co-infected with tuberculosis (TB), particularly in areas where these conditions are common. We describe the clinical features, laboratory findings and outcome of hospitalised PWH and human immunodeficiency virus (HIV)-uninfected COVID-19 patients as well as those co-infected with tuberculosis (TB).MethodsWe conducted a multicentre cohort study across three hospitals in Cape Town, South Africa. All adults requiring hospitalisation with confirmed COVID-19 pneumonia from March to July 2020 were analysed.ResultsPWH comprised 270 (19%) of 1434 admissions. There were 47 patients with active tuberculosis (3.3%), of whom 29 (62%) were PWH. Three-hundred and seventy-three patients (26%) died. The mortality in PWH (n = 71, 26%) and HIV-uninfected patients (n = 296, 25%) was comparable. In patients with TB, PWH had a higher mortality than HIV-uninfected patients (n = 11, 38% vs n = 3, 20%; p = 0.001). In multivariable survival analysis a higher risk of death was associated with older age (Adjusted Hazard Ratio (AHR) 1.03 95%CI 1.02–1.03, p < 0.001), male sex (AHR1.38 (95%CI 1.12–1.72, p = 0.003) and being “overweight or obese” (AHR 1.30 95%CI 1.03–1.61 p = 0.024). HIV (AHR 1.28 95%CI 0.95–1.72, p 0.11) and active TB (AHR 1.50 95%CI 0.84–2.67, p = 0.17) were not independently associated with increased risk of COVID-19 death. Risk factors for inpatient mortality in PWH included CD4 cell count < 200 cells/mm3, higher admission oxygen requirements, absolute white cell counts, neutrophil/lymphocyte ratios, C-reactive protein, and creatinine levels.ConclusionIn a population with high prevalence of HIV and TB, being overweight/obese was associated with increased risk of mortality in COVID-19 hospital admissions, emphasising the need for public health interventions in this patient population.

Introduction: Women living with human immunodeficiency virus (WLHIV) may face additional challenges and differential birth outcomes when compared with women without human immunodeficiency virus (HIV). There is limited research to date studying birth outcomes among a nationally representative sample of WLHIV. This study compares stillbirth and prematurity rates between HIV-positive (HIV+) and HIV-negative (HIV−) mothers in the Medicaid program.Methods: We used 12 years (2001–2012) of Medicaid Analytic eXtract data. We included Medicaid claims from the 14 states with the highest prevalence of HIV: California, Florida, Georgia, Illinois, Louisiana, Massachusetts, Maryland, North Carolina, New Jersey, New York, Ohio, Pennsylvania, Texas, and Virginia. Primary outcomes were stillbirth and preterm birth. We used logistic regression models adjusting for age, race, Medicaid coverage, eligibility, substance use, rurality, comorbidities, and state fixed effects to compare differences in rates for women with and without HIV.Results: Our study included a total of 33,078 HIV+ and 7,663,758 HIV− pregnancies from Medicaid enrollees between 2001 and 2012. The proportions of stillbirths and preterm births were higher for HIV+ when compared with HIV− mothers (0.9% vs. 0.7% and 8.0% vs. 6.6%, p < 0.0001). After adjusting for covariates, being HIV+ was not significantly associated with both stillbirth (odds ratio [OR]: 1.05) or prematurity (OR: 1.01). Black race was a strong independent predictor of both stillbirth and prematurity (OR: 1.99 and 1.51, p < 0.01). Rurality and substance abuse were not associated with either outcome.Conclusions: After adjustment for relevant covariates, maternal HIV infection was not associated with increased rates of stillbirth or preterm birth in the Medicaid program in the United States. It is imperative that we understand and eliminate the clinical, social, and contextual factors that are responsible for the strong association between black race and poor perinatal outcomes that we observe.

3D ultrasound (US) has become prevalent due to its rich spatial and diagnostic information not contained in 2D US. Moreover, 3D US can contain multiple standard planes (SPs) in one shot. Thus, automatically localizing SPs in 3D US has the potential to improve user-independence and scanning-efficiency. However, manual SP localization in 3D US is challenging because of the low image quality, huge search space and large anatomical variability. In this work, we propose a novel multi-agent reinforcement learning (MARL) framework to simultaneously localize multiple SPs in 3D US. Our contribution is four-fold. First, our proposed method is general and it can accurately localize multiple SPs in different challenging US datasets. Second, we equip the MARL system with a recurrent neural network (RNN) based collaborative module, which can strengthen the communication among agents and learn the spatial relationship among planes effectively. Third, we explore to adopt the neural architecture search (NAS) to automatically design the network architecture of both the agents and the collaborative module. Last, we believe we are the first to realize automatic SP localization in pelvic US volumes, and note that our approach can handle both normal and abnormal uterus cases. Extensively validated on two challenging datasets of the uterus and fetal brain, our proposed method achieves the average localization accuracy of 7.03 degrees/1.59mm and 9.75 degrees/1.19mm. Experimental results show that our light-weight MARL model has higher accuracy than state-of-the-art methods.