Reinforcement learning (RL) solutions have shown considerable promise for automating the defence of networks to cyber attacks. However, a limitation to their real world deployment is the sample efficiency and inflexibility of RL agents. This means that even small changes to attack types requires a new agent to be trained from scratch. Meta-learning for RL aims to improve the sample efficiency of training agents by encoding pre-training information that assists fast adaptation. This work focusses on two key meta-learning approaches, MAML and ML3, representing differing approaches to encoding meta learning knowledge. Both approaches are limited to sets of environments that use the same action and observation space. To overcome this we also present an extension to ML3, Gen ML3 that removes this requirement by training the learned loss on the reward information only. Experiments have been conducted on a distribution of network setups based on the PrimAITE environment. All approaches demonstrated improvements in sample efficiency against a PPO baseline for a range of automated cyber defence (ACD) tasks. We also show effective meta-learning across network topologies with Gen ML3.

Background: Few studies have evaluated COVID-19 vaccine effectiveness (VE) in middle-income countries, particularly in eastern Europe. We aimed to estimate COVID-19 VE against SARS-CoV-2-confirmed hospitalizations and severe outcomes in Kosovo. Methods: We conducted a test-negative case-control study using data from Kosovo’s severe acute respiratory infection (SARI) sentinel surveillance system from January 2022 to June 2024. We enrolled adult patients aged ≥18 years hospitalized with SARI. From all patients, we collected clinical data, vaccination history and a nasopharyngeal specimen, which was tested for SARS-CoV-2 using RT-PCR. SARS-CoV-2-positive patients were cases; those testing negative were controls. We estimated VE overall and against severe outcomes (requiring oxygen, intensive care admission, or in-hospital death) using logistic regression, adjusting for age, sex, and comorbidities, calculating VE as (1–adjusted odds ratio)×100. Results: We included 564 SARI patients; 218 (39%) tested positive for SARS-CoV-2. Overall, 24% of SARI patients had received at least one COVID-19 vaccine dose in the previous 12 months. VE against SARS-CoV-2-confirmed SARI hospitalization among all adults was 75% (95%CI:37–90%) at 14–179 days post-vaccination, and 24% (95% CI:-35–58%) at 180–364 days. In adults ≥60 years, VE was 59% (95%CI:-10–84%) at 14–179 days post-vaccination, and -37% (95%CI:-188–5%) at 180–364 days. VE against severe outcomes was 70% (95% CI:-2–91%) at 14–179 days, and 8% (95%CI:-128–63%) at 180–364 days. Conclusions: In Kosovo, COVID-19 vaccination offered significant protection against hospitalization and severe outcomes. Effectiveness waned after 6 months, highlighting the need for periodic booster doses.

Gout is a metabolic disorder characterized by the accumulation of monosodium urate (MSU) crystals in joints, leading to acute inflammatory arthritis. The gout signaling pathway is activated by the deposition of these crystals, which are recognized by pattern recognition receptors (PRRs) such as Toll-like receptors (TLRs) and NOD-like receptors (NLRs), particularly the NLRP3 inflammasome. This activation triggers the release of pro-inflammatory cytokines, including interleukin-1β (IL-1β), tumor necrosis factor-alpha (TNF-α), and interleukin-6 (IL-6), which mediate the inflammatory response. The pathway also involves the recruitment of neutrophils and other immune cells, perpetuating joint inflammation. The document further highlights the role of NF-κB signaling and caspase-1 activation in cytokine maturation. Understanding the molecular mechanisms underlying this pathway offers potential targets for therapeutic intervention, emphasizing the need for novel anti-inflammatory agents to manage gout effectively.

Background : Non-suicidal self-injury (NSSI) is primarily performed for negative reinforcement, which may be accomplished by facilitating avoidance of negative affect. This study aimed to identify whether an aversion to cognitive effort in the form of effort discounting underlies NSSI and other avoidant emotion regulation strategies. Methods : Adults (N= 149), 52% with a history of NSSI, completed a behavioral measure of effort discounting, the preference to expend less effort despite attaining less relief from an aversive experience. We used exploratory factor analysis to identify a dimension of avoidant vs. engagement-oriented affect regulation strategies, and we extracted factor scores from this dimension. Participants self-reported their past engagement in NSSI, perception of the effort required by non-NSSI affect regulation strategies, and past use of such strategies. Results : NSSI frequency was associated with the perception that other, non-NSSI affect regulation strategies were more effortful and effort discounting was associated with the use of more avoidant coping strategies. We did not find an association between avoidant affect regulation strategy use and NSSI frequency, regardless of whether NSSI was used for emotion regulation. Conclusions : Our results provide preliminary support that avoidant affect regulation strategies, possibly including NSSI, could be driven by exaggerated effort discounting and the perception that avoidance requires less cognitive effort.

Speciation – the process by which new species arise – is fundamentally influenced by population-level factors such as population size and demographic dynamics. Here, we review how population size and its dynamics shape speciation mechanisms and the generation of biodiversity across evolutionary scales. Small populations can undergo rapid genetic changes via drift and founder events, potentially promoting speciation, whereas large populations harbour greater genetic diversity and adaptive potential, influencing divergence in different ways. We examine how genetic drift and gene flow interact to facilitate or impede speciation under various scenarios and how extinction risk during adaptation affects the development of reproductive isolation. Finally, we explore how speciation rates and the persistence of populations/species over time are related in theoretical models. We highlight insights from mathematical models – especially those explicitly incorporating population size – and identify open questions for future research.

This short communication develops a variational formulation for modeling a silicon semiconductor design utilizing phosphorus atoms as a dopant substance. The results are based on standard tools of calculus of variations and optimization theory.

This article develops a variational formulation for modeling a protium hydrogen molecular ionization obtained through a high temperature scalar field and an appropriate electric one action. The results are based on standard tools of calculus of variations and optimization theory. Finally, we highlight the context here addressed is essentially an Euler-Bernoullian one and includes the establishment of a new approximate Bernoulli-perfect-gas type equation.

WJP was responsible for scientific background, formulation of the research question, conceptual manuscript design, figure development, and drafting of the manuscript inclusive reference management. LT and DY contributed to figure conceptualization and manuscript editing and revision. OH, YM, ML, and TH were involved in patient recruitment and data collection, as well as manuscript editing and revision. KM was responsible for figure design and adaptation, manuscript drafting, editing, and revision.

Introducción a Machine Learning Tiempo Sugerido: 50 minutosDescripción: En esta lección, aprenderemos los conceptos básicos de Machine Learning. A través de una actividad práctica con la herramienta de Google Teacheable Machine, los estudiantes descubriran como las computadoras pueden "aprender a aprender". Entrenaran un modelo simple de clasificación de objetos por fotografía y luego reflexionaran como se aplican estos conocimientos en el mundo real (reconocimiento facial, ads de recomendaciones, etc.).Vocabulario:Machine learningDatosAlgoritmoEntrenamientoClasificaciónPredicciónModeloVariables independientes (input)Variable dependiente (output)Objetivos:Comprender conceptos básicos de machine learning.Entrenar un modelo sencillo con diferentes categorías.Identificar aplicaciones cotidianas de ML.Materiales de proyecto necesarios:Computadoras o tabletas con acceso a internetCámara webAcceso a Teacheable MachinePapel, lápiz o cuaderno para tomar notasCartulinas de coloresTijerasMarcadorRecursos multimedia:¿Qué es el aprendizaje automatizado? Vídeo de Youtube.Presentación de apoyoPlan de lección:Parte 1: Introducción a Machine Learning (15 min)1. Activación de conocimientos previosPregunta razonadora: ¿Qué crees que significa que una máquina aprenda?Ejemplos cotidianos: Recomendaciones de Netflix, filtros de spam, carros inteligentes.2. ¿Qué es el Machine Learning?Explicación: ML es una rama de la inteligencia aritficial donde las computadoras aprenden de datos para tomar decisiones o hacer predicciones.3. Un poco de mátematicas Los modelos de ML usan funciones matemáticas para encontrar patrones. Por ejemplo:        Una recta y = mx + b, puede separar datos. Los modelos se entrenan por medio del ajuste de         parámetros para minimizar el error (usando la función como el error cuadrático medio).4. ¿Quiénes trabajan en Machine Learning? Presentar a los ingenieros responsables del MLIngenieros de Machine LearningIngenieros de SoftwareIngenieros de Inteligencia ArtificialParte 2: ¡Entrena tu propio modelo! (25 min) Dividir a lo estudiantes en grupos pequeñosLos estudiantes utilizaran las cartulinas de colores, las tijeras y los marcadores para crear diferentes formas que son las que su modelo reconocerá y clasificará.Cada grupo ingresará a Teacheable Machine.Crean un modelo de clasificación de imágenes. Por ejemplo, "circulo rojo vs. cuadrado verde".Capturan los ejemplos con su cámara y entrenan el modelo.Prueban el modelo en vivo y documentan sus observaciones.Parte 3: Discusión y reflexión (10 min)¿Qué funcionó bien? ¿Qué no?¿Qué pasaría si entrenamos con imágenes borrosas o repetidas?¿Qué pasa si sólo entrenamos el mismo tamaño de figuras?Reflexión: ¿Dónde ves ML en tu vida diaria?

G protein-coupled receptors (GPCRs) are central regulators of neurotransmission and neuroplasticity, influencing the pharmacodynamics of psychoactive substances. As key nodes in signal transduction, they modulate second messenger systems, ion channel activity, and synaptic communication. Despite extensive research on GPCRs as therapeutic targets, their specific role in addiction-related neuroadaptations remains insufficiently understood. This review systematically examines how core GPCR subtypes; dopamine, opioid, cannabinoid, and metabotropic glutamate receptors mediate the reinforcing properties of addictive drugs. It highlights the dual roles of GPCR-dependent and independent pathways in altering reward circuitry and driving compulsive drug-seeking. The review also explores recent innovations in structure-based drug design, including biased agonism, allosteric modulators, and GPCR conformational dynamics, for next-generation addiction therapies. By integrating molecular pharmacology with addiction neuroscience, this review identifies translational challenges and opportunities, underscoring the value of GPCR-targeted pharmacotherapies. Ultimately, it advocates for interdisciplinary strategies to bridge preclinical findings with effective clinical interventions for substance use disorders.

We analyze a system of diffuse competition in an ant community in a coffee agroecosystem. The community is constrained by space since the dominant competitor (Azteca sericeasur) has a limited range of influence within which it excludes other species, but outside that range, the other species may persist. We use a simple dynamic model in which variable A, the dominant competitor, competes strongly with variable S, the collection of all other species in the system, the latter acting as a strong competitor in the context of diffuse competition. Two qualitatively distinct outcomes are expected depending on the strength of competition: 1) a simple monotonic increase in species richness as distance from the Azteca nest increases, or 2) a range of distance from the nest in which there is bi-stability, bookended by critical transitions. Studying five nests in wet and dry seasons over a two-year period, we find evidence for bi-stability.

Objective: To evaluate the associations between Listeria monocytogenes exposure, infant perinatal outcomes and hospital resource use. Design: Cross-sectional analysis. Setting: Antenatal clinics in Australia. Population: A cohort of 1,604 Australian mother-infant dyads. Methods: Multinomial, negative binomial, and hurdle models, were performed to examine associations between Listeria monocytogenes exposure with infant outcomes and hospital resource use. All models were adjusted for covariates including smoking, parity, maternal age, BMI, and SEIFA IRSAD decile. Main outcomes measures: Maternal L. monocytogenes exposure was estimated from self-reported intake of foods from a validated food frequency questionnaire that potentially harbour L. monocytogenes. Infant outcomes obtained from hospital medical records included birth mode, preterm birth, birthweight, and admission to special care nursery (SCN)/neonatal intensive care unit (NICU). Infant hospital resource use was measured by infant length of stay (LOS), LOS in SCN/NICU, and days excluding SCN/NICU. Results: Mean (SD) maternal age was 32.0 (5.0) years, and median (IQR) gestation was 39.0 (38.1, 40.0) weeks. While adjusted results showed a statistically significant association between LFES and reduced infant LOS excluding SCN/NICU (β= 0.99; 95%CI 0.979, 0.998, p<0.03), the effect size was minimal, with minor clinical significance. There were no significant associations with infant birth mode, preterm birth, low birthweight, size for gestational age, macrosomia, admission to NICU/SCN, total LOS in hospital, and SCN/NICU (all p>0.05). Conclusions: Future research should explore these associations among ethnically diverse women at earlier stage of pregnancy and include the assessment of food safety practices in the analyses.

Introduction: Laparoscopic lateral suspension (LLS) with mesh is indicated for the treatment of apical prolapse. However, clinical data of this technique still remain limited. Objective: To evaluate the clinical outcomes and changes in the vaginal axis of LLS with mesh for apical and anterior prolapse. Design: Retrospective single center clinical study. Setting: Conducted at the department of Obstetrics and Gynecology, West China Second University Hospital, China. Sample: Ninety-seven patients underwent LLS between 2020 and 2024. Methods: Retrospective analysis. Main Outcome Measures: Anatomical success rate, subjective cure rate, complications rate and changes in the vaginal axis before and after LLS by magnetic resonance imaging (MRI). Results: The anatomical success rate was 98.97% for apical prolapse and 85.57% for anterior prolapse, with a perioperative complication rate of 3.09% in the two-year follow-up. Mesh exposure at the vaginal cuff was observed in five patients (5.15%), all of whom had a concomitant hysterectomy. The patient satisfactory rate (evaluated by PGI-I score ≤3) was 96.67% at 24 months postoperatively. On MRI, no significant differences were observed between the Post-LLS and nulliparous women regarding the angle between the lower and upper vagina or the Cx/Apical line at rest (p> 0.05). Conclusion: LLS is an effective and safe surgical technique for treatment of apical prolapse with or without anterior prolapse, demonstrating low complication rates and high anatomic curative rates in short-term follow-up. MRI examination showed that a level-Ⅰ defect can be effectively addressed well through LLS, highlighting its capability to restore apical support.

Background Ascending placentitis is an important cause of fetal mortality and neonatal morbidity in the horse. While various diagnostic and treatment protocols exist for placentitis, there is a lack of data investigating what options veterinary surgeons are selecting for case work-up and management. Objectives To identify the diagnostic and treatment options for ascending placentitis utilised by veterinary surgeons in the UK and Ireland, and to ascertain what influences the selection of these options. Study design Observational study: online survey Methods An online questionnaire was distributed to veterinary surgeons in the UK and Ireland via email and social media. The questionnaire consisted of 17 questions asking about participant demographics, specifics on case work-up including diagnostic and treatment selections, as well as the rationale behind selections. Results A total of 80 veterinary surgeons participated in the survey, with 74 completing it in full. Clinical signs were often used to diagnose placentitis (98.7%), followed by CTUP and rectal ultrasound (87% and 75.3%, respectively). Treatment of placentitis typically involved a combination of antibiotics, anti-inflammatories and progesterone (80.3%). Trimethoprim-sulfonamide was the most frequently selected antibiotic (100%). Flunixin and phenylbutazone were the most popular anti-inflammatory choices and were used equally frequently (60.3% each). Progesterone was used by 90.8% of participants. Owner finances limited the selection of diagnostics and treatments in many cases (71.1% and 74.6% respectively). Main limitations Risk of selection bias. Opportunistic sampling. Conclusions Veterinary surgeons in the UK and Ireland are generally using appropriate choices for diagnostics and treatments, in keeping with the current literature. There is a need for expansion of CPD and undergraduate training to improve the confidence and competence of veterinary surgeons treating equines for ascending placentitis. Finances are a major limiting factor for choices of diagnostics and lengths of treatment.

Introduction: Lung cancer is a leading cause of cancer-related deaths, with an annual mortality rate of nearly 1.8 million. Among its subtypes, lung adenocarcinoma is increasingly prevalent. Due to its often asymptomatic nature, most patients are diagnosed at advanced stages, missing the optimal window for surgery and resulting in a poor 5-year survival rate. Case presentation: We present the case of an 83-year-old patient with stage IV lung adenocarcinoma treated with four cycles of intravenous Tirelizumab and two cycles of intratumoral Tirelizumab via tracheoscopy. The patient achieved partial clinical remission, with tumor lesions continuing to shrink upon follow-up by August 24, 2024. No treatment-related adverse reactions were observed, and the patient’s immune function remained normal. Conclusions: This case suggests that systemic intravenous Tirelizumab combined with tracheoscopic intratumoral injection may offer a safe and effective treatment strategy for lung adenocarcinoma, particularly for patients unsuitable for chemotherapy or surgery. However, further prospective studies and clinical trials are needed to validate these findings.

Zepeng Zheng, Shilian Xu, Xiaomiao Zeng, Xiaojun Chen, Xiaoli Liu, Shuai Fu, Shuning ZhangDepartment of Obstetrical, Shenshan Medical Central, Memorial Hospital of Sun Yat-sen University, Shanwei 516600, ChinaCorrespondence：Shuai Fu and Shuning Zhang are co-corresponding authors.

Objective: To assess the dynamics in blood concentrations of vitamins (A, B6, B12, D, E,), trace elements such as selenium, magnesium, zinc, and iron (transferrin), and metabolite homocysteine during pregnancy and postpartum. Design: Cross-sectional, national cohort study conducted between January and June 2024. Setting: Slovakia. Population: Pregnant and postpartum women. Methods: From venous blood and capillary dry blood spot micronutrients were analysed using standard biochemical and biophysical methods. Main Outcome Measures: Temporal changes in blood micronutrient levels across pregnancy and postpartum. Results: Our findings demonstrate significant temporal variations in maternal micronutrient levels. While some nutrients remained stable (vitamin D, folate), others showed a decline (vitamin A, B12, iron, zinc), whereas certain nutrients increased (vitamin E) during pregnancy. Vitamin E concentrations in the 3 rd trimester frequently exceeded reference values for the general adult population, whereas zinc levels significantly declined postpartum. We observed high prevalence of vitamin B12 and iron deficiencies, as indicated by transferrin saturation, particularly in the 3 rd trimester. Vitamin D deficiency was prevalent throughout pregnancy and postpartum. Finally, our analysis demonstrated that dried blood spot (DBS) technology provides comparable results to venous blood analysis for measuring vitamin A, D and homocysteine levels. Conclusion: These findings highlight the dynamic nature of maternal micronutrient status, emphasizing the necessity of systematic nutritional surveillance during pregnancy and postpartum. The integration of DBS technology into routine clinical practice could facilitate early detection and management of maternal nutrient deficiencies.

Closely-related species living in sympatry are often partitioned into divergent ecological niches. Such specialization can be enabled by the evolution of divergent traits enhancing adaptation to different niches. In this study, we investigate the partitioning of closely-related butterfly species into different forest strata and daily activity time and test the effects of such spatio-temporal niches on the evolution of thermal traits. First, using experiments in the field in Amazonia, we precisely characterized the daily activity patterns of nine species of Morpho butterflies, therefore documenting extensive temporal segregation among species and observing significant variations in temperature between their respective niches. Using controlled experiments in the lab, we then tested the thermal tolerance of wild individuals to both hot and cold conditions. The vertical distribution of species (understory vs. canopy micro-habitats) had a significant effect on several thermal traits, even when controlling for the phylogenetic distances between species, suggesting that forest stratification may shaped thermal adaptation in these tropical butterflies. However, butterfly activity time did not correlate with any thermal traits measured. The extensive temporal segregation observed between these sympatric species might thus stem from ecological interactions observed between species rather than thermal factors.

Bacterial adaptation to environmental stress is greatly aided by small regulatory RNAs (sRNAs). Here, we examine metabolic adaptation to ciprofloxacin (CIP) stress via sRNA in Klebsiella sp. SG01. We identified intergenic regions enriched with sRNAs and expression patterns described by examining transcriptomic datasets. Nine potential sRNAs, glmZ, sgrS, gcvB, spot42, micA, istR, sraL, icsR, and fnrS were identified by combining differential expression analysis with established computational pipeline. Furthermore, an FMN aptamer (RFN element) was found, indicating a possible function in regulation mediated by riboswitches. These sRNAs target important metabolic pathways, such as the tricarboxylic acid cycle (TCA), redox homeostasis, and the transport of sugars and amino acids, according to functional annotation using TargetRNA3 predictions and literature mining. GcvB was found to negatively regulate ABC transporters and amino acid permeases, whereas Spot42 was found to regulate sugar transporters and TCA intermediates. A dynamic rewiring of the sRNA network under CIP stress was also suggested by transcriptomic analysis, which showed a regulatory shift marked by the downregulation of Hfq and the upregulation of ProQ and CsrA. This study opens the door for further investigation into the complex post-transcriptional regulatory systems that bacteria use to maximize resource allocation and stress resilience.

Background: Currently, there are more studies on the increase of bile acids during pregnancy, but less attention has been paid to the decrease of bile acids during pregnancy. Research aim: To explore the impact of low bile acid levels during the third trimester of pregnancy on pregnancy and neonatal outcomes. Methods: 429 pregnant women had total bile acid levels of less than 1 µmmol/L in both tests, and these were selected as the study group. 501 pregnant women with normal total bile acid levels in both tests were randomly selected from them as the control group. Results: Compared with the normal control group, the lower bile acid levels were associated with higher age(30.64±4.43 vs 29.27±3.63 years; P<0.001) and smaller gestational age(38.73±1.29 vs 39.34±0.97 weeks; P<0.001). There were significant differences between the two groups in the incidences of gestational diabetes, hypothyroidism and premature rupture of membranes. These conditions were more common in pregnant women with low bile acid levels. Conclusion: Pregnant women with low bile acid levels have a higher incidence of gestational diabetes, hypothyroidism and premature rupture of membranes.

A document by Dr. Tamil Bharathan T K. Click on the document to view its contents.

The acid-controlled single-component retro-aldol/Michael addition cascade reaction and [4+2] cycloaddition of benzofuran-derived azadienes (BDAs) are reported for the first time. Under the conditions of trifluoromethanesulfonic acid as the catalyst and with the addition of water, BDAs initiate the retro-aldol reaction, followed by a 1,4-Michael addition, yielding (arylmethylene)bis-(dibenzofuran) products with excellent yields and broad substrate applicability. This represents the first application of BDAs in a retro-aldol reaction. In contrast, in the absence of water and with boron trifluoride etherate as the catalyst, BDAs undergo a [4+2] cycloaddition reaction, constructing the spiro[benzofuran-2,3’-benzofuro[3,2-b]pyridine] framework with high yields and diastereoselectivity. The method features mild conditions, high atom economy, and provides a new approach for constructing benzofuran scaffold derivatives.

Here, we report the first structural characterization of lipid A from Pragia fontium isolated from well water in South Bohemia, Czech Republic, in 1982 [(Aldova et al., 1988)](#ref-0001), using our Fast Lipid Analysis Technique (FLAT) and its tandem mass spectrometry version, FLAT n, [(Sorensen et al., 2020)](#ref-0009) [(Yang et al., 2022)](#ref-0011). This analysis identified the major P. fontium lipid A molecule, distinguishing it from other Enterobacteriaceae. Lipid A is the bioactive component of lipopolysaccharide (LPS), essential for the structural integrity of Gram-negative bacterial outer membranes and a key modulator of host immune responses. While the lipid A structures of many pathogenic bacteria have been extensively studied, limited information is available on non-pathogenic or environmental isolates, such as P. fontium. These findings provide new insights into the structural biology of this organism and may have implications for its interaction with the environment and host immune responses.

Model Predictive Control (MPC) is an advanced control method that is gaining interest thanks to its ability to control multivariable systems and handle constraints. Nevertheless, the effectiveness of MPC depends on the accuracy of the mathematical model that is used to make predictions about the system state trajectory. As the design of an accurate first-principle model of the plant to be controlled can be challenging, an increasing number of predictive controllers based on black-box models have been proposed. Among the black-box models that have been considered, Gaussian Processes (GPs) have been of particular relevance. However, the implications of GP model uncertainties on the recursive feasibility of predictive controllers and the stability of the closed-loop system are still a debated issue. In this manuscript, we resort to recent results about the stabilizing properties of robust MPC algorithms based on the contractivity of the system to be controlled, to build a robust MPC based on GPs that can guarantee stability and recursive feasibility under mild assumptions.