The integration of artificial intelligence into healthcare and dental practice remains pivotal for overcoming some of the most persistent challenges associated with fragmented records, resource constraints, and variable quality of care. This paper presents the CARE Framework, an integrated AI-powered platform developed to digitize, automate, and optimize patient management across medical and dental domains. It integrates a robust electronic medical record infrastructure, advanced machine learning modules for clinical decision support and predictive analytics, and workflow automation, guaranteeing interoperability via HL7 and FHIR standards. A mixed-methods evaluation was conducted across 100,000 patients with diverse medical and dental profiles. The quantitative outcomes showed significant improvements after a 12-month CARE implementation, with hospital readmission rates decreasing from 16.7
Antimicrobial resistance (AMR) is a mounting public health crisis threatening the efficacy of modern antibiotics worldwide. Artificial intelligence (AI) is poised to transform AMR management through innovations in diagnostics, drug discovery, clinical stewardship, and surveillance. AI-powered diagnostic tools analyze genomic, imaging, and clinical data to enable rapid and accurate detection of resistant pathogens, reducing time to targeted treatment and improving patient outcomes. In drug discovery, AI accelerates the identification of novel antibiotics and optimal candidates for repurposing by computationally screening vast chemical libraries and predicting antimicrobial activity, resulting in breakthroughs against multi-drug-resistant bacteria. Clinical decision support systems leverage AI to optimize antibiotic prescribing, minimize inappropriate use, and enhance antimicrobial stewardship. On a global scale, AI-driven surveillance and national drug registries provide real-time insights into resistance trends, supporting targeted interventions and regulatory compliance. Recent advances, including generative AI-designed antibiotics and integrative public health platforms, illustrate the profound potential of AI to curb resistance, extend the utility of existing antimicrobials, and guide policy. Continued interdisciplinary collaboration and robust data governance will be essential for realizing AI’s promise in reversing AMR and safeguarding future infectious disease control.
It is important that tree transpirationTranspiration rates can be predicted accurately because without it we cannot fully understand forest hydrology, manage water resources, or quantify climate effects. Traditional models typically struggle with the nonlinear dynamic nature of transpirationTranspiration. In this study, we evaluate three long short-term memory (LSTM)Long Short-Term Memory (LSTM)-based architectures: Vanilla LSTM, Stacked LSTM, and Bidirectional LSTM (BiLSTM) to determine a robust baseline model for predicting tree transpirationTranspiration. The measurements of the sap flow sensor and meteorological data were used to test each model in a five-fold cross-validation approach and were evaluated based on the coefficient of determination ( R^2 ), mean absolute error (MAE), root mean square error (RMSE), and statistical tests. Our results indicate that the BiLSTM model outperformed the other architectures, quantifying the highest accuracy and lowest error rates. Motivated by establishing a baseline, we propose a modified BiLSTM model, consisting of Conv1D layers, and is designed to improve feature extraction. Our hybrid architecture outperformed the BiLSTM baseline, with an average R^2 of 0.987, and had a lower average MAE and RMSE. These findings highlight the importance of the architecture used in environmental time-series modeling and offer a new way to improve transpirationTranspiration forecastingForecasting using both deep temporal and convolutional features.
As the demand for high-performance computingHigh-performance computing systems (HPC) continues to grow across various domains, from scientific research to industrial innovation, they enable complex simulations, data analysis, and modelling tasks at unprecedented speeds. However, the significant energy consumption associated with such powerful computing capabilities presents a major challenge, particularly concerning environmental impact in today’s world. Therefore, novel strategies are needed to dynamically manage system workload to optimize energy consumption, enhance sustainabilitySustainability, and minimize environmental impact. This study addresses the energy consumption challenge in high-performance computingHigh-performance computing by proposing an improved bio-inspired algorithm for energy-efficient dynamic scheduling. The algorithm utilizes OpenMP parallel programming directives to distribute the system workload across multiple processors dynamically. It integrates optimization techniques to maximize local exploitation and global exploration within the problem search space. The GEM5 simulator is used to evaluate the performance and energy consumption of the proposed approach. Results show that the proposed algorithm outperforms non-bio-inspired approaches, achieving up to 98
