We introduce PandemIQ Llama, a domain-adapted large language model (LLM) designed specifically for pandemic intelligence applications. Building on the pre-trained Llama-3.1-8B model, we conducted continuous training using our curated Pandemic Corpus. This dataset was assembled from authoritative public health sources, scientific literature, and specialized knowledge repositories, comprising 508,924 documents totaling 5.8 billion tokens, which is the largest pandemic domain specific data cohort for LLM training. Benefited from our curated large data cohorts and through continuous training leveraging extensive computational resources, the developed PandemIQ Llama model can extract critical domain knowledge on pandemic, which is typically underrepresented in general-purpose language models, To evaluate its performance, we conducted comprehensive comparison of PandemIQ Llama with both prompt-engineered and task-specific fine-tuned baseline models using two tasks: the Biomedical Alert News Question Answering task (1,508 reports with 30 expert-generated questions each) and the Disease Event Type Classification benchmark (4,500 news snippets across eight disease categories). PandemIQ Llama demonstrated substantial improvements over strong baseline models, achieving performance gains ranging from 3.8% to 10.97%. These results suggest that PandemIQ Llama could significantly enhance public health surveillance and analysis capabilities. In addition, our result also suggests that the LLMs can perform better with continuous training than fine-tuning on domain specific tasks. Social Impact: The BEACON platform, powered by our model, launched and now serves over 100 government and multilateral public health organizations and users across 154 countries. Analytics from the platform is being integrated into the Epidemic Intelligence from Open Sources system run by the World Health Organization. This integration will provide public health decision-makers with a powerful LLM-based tool for pandemic surveillance.

During the COVID-19 pandemic, communities faced two conflicting objectives: 1) minimizing infections among vulnerable populations with higher risk for severe illness and 2) enabling reopening to revive American livelihoods. The U.S. pandemic strategy myopically considered one objective at a time, with lockdowns that addressed the former, but was detrimental to the latter, and phased reopening that pursued the latter, but lost control over the former. How could we prioritize interventions to simultaneously minimize cases of severe illness and fatalities while reopening? A team of researchers anchored by the Center on Stochastic Modeling, Optimization, & Statistics (COSMOS), The University of Texas at Arlington, has formulated a computationally efficient optimization framework, referred to as COSMOS COVID-19 Linear Programming (CC19LP), to study the delicate balance between the expected fatality rate due to cases of severe illness and the level of normalcy in the community. The key to the CC19LP framework is a focus on “key contacts” that separate individuals at higher risk from the rest of the population. CC19LP minimizes expected fatalities by optimizing the use of available interventions, namely, COVID-19 testing, personal protective equipment (PPE), COVID-19 vaccines, and social precautions, such as distancing, handwashing, and face coverings. A C3.ai award-winning online CC19LP tool is accessible from the COSMOS COVID-19 project site (https://cosmos.uta.edu/projects/covid-19/) and has been tested for all 3142 U.S. county areas. Results are demonstrated for several metropolitan counties with a deeper investigation for Miami-Dade County in Florida.Note to Practitioners—In this article, a computationally fast optimization framework is presented to study the delicate balance between reopening U.S. communities and controlling severe cases of COVID-19 that lead to hospitalizations and fatalities. This framework can provide guidance to decision-makers on optimal intervention strategies for protecting high-risk individuals while reopening communities. This optimization framework demonstrates a practical approach to conduct decision-making in an uncertain environment and can be useful for the prioritization of resources and interventions in the case of future epidemics or pandemics. Resources on understanding and implementing the framework are publicly available, including an award-winning online optimization tool that automatically accesses county-level data from Census, Centers for Disease Control and Prevention (CDC), and Johns Hopkins COVID-19 repositories.

The investigation of risk factors associated with hypertension patients has been extensively studied in the past decades. However, the pattern of natural progressive trajectories to hypertension from nonhypertensive states was rarely explored. In this study, we are interested in discovering the underlying transition patterns between different blood pressure states, namely normal state, elevated state, and hypertensive state among the working population in the United States. A multi-state Markov model was built based on 88,966 clinical records from 34,719 participants we collected during the worksite preventive screening from 2012 to 2018. We first investigated the various risk factors, and we found that body mass index (BMI) is the most critical factor for developing new-onset hypertension. The transition probabilities, survival probabilities, and sojourn time of each state were derived given different levels of BMI, age groups, and gender categories. We found the underweight participants are more likely to remain in the current nonhypertensive states within 3 years, while extremely obese participants have a higher probability of developing hypertension. We discovered the distinct transition patterns among male and female participants. On average, the sojourn time in the normal state for normal-weight participants is 4.33 years for females and 2.18 years for their male counterparts. For the extremely obese participants, the average sojourn time in the normal state is 1.38 years for females and 0.71 years for males. In the end, a web-based graphical user interface (GUI) application was developed for clinicians to visualize the impact of behavioral interventions on delaying the progression of hypertension. Our analysis can provide a unique insight into hypertension research and proactive interventions.

In this paper, the stratospheric aerosol extinction profiles are retrieved from Scanning Imaging Absorption Spectrometer for Atmospheric Cartography (SCIAMACHY) limb scatter measurements. The retrieved aerosol profiles are validated by comparison with the collocated Stratospheric Aerosol and Gas Experiment (SAGE) II solar occultation measurements. The comparisons are performed for three latitudinal bands (20°-30°N, 30°-40°N, 40°-50°N). In the 15-30 km altitude range, mean relative differences between the SCIAMACHY limb retrievals and the SAGE II measurements are found to be generally within 20%. The mean column aerosol optical depths in this altitude range agree within 3 %. A validated dataset of aerosol extinction profiles from SCIAMACHY limb measurements is expected to facilitate the understanding of physical and chemical processes in the stratosphere.

This paper is devoted to retrieve the stratospheric aerosol extinction profiles from the limb measurememts performed by Scanning Imaging Absorption Spectrometer for Atmospheric Cartography (SCIAMACHY) instrument on the European Environmental Satellite. The SCIATRAN radiative transfer model is employed to simulate the limb radiance received by the SCIAMACHY instrument, and an optimal estimation method is used to retrieve the aerosol extinction profiles. For this retrieval, the measurement vector is defined as the ratio of two normalized spectral channels in order to minimize the effects of the uncertainty parameters and to reduce the impact of instrument calibration errors. The retrieved aerosol extinction profiles are compared with the Stratospheric Aerosol and Gas Experiment (SAGE) II measurements, and the results show that for the zonal mean profiles, SCIAMACHY retrievals exhibit good agreement with SAGE II measurements, with the relative differences being generally within 20% over the altitude range of 12-28 km.

For adaptive distributed applications, the computational complexity associated with each computational region varies continuously and dramatically both in space and time throughout the life cycle of the application execution. Consequently, static scheduling techniques are inefficient for such applications. In this paper, we present an autonomic runtime manager (ARM) that uses the application spatial and temporal characteristics as well as resource status as the main criteria to self optimize the execution of distributed applications at runtime. We applied the ARM system to a wildfire simulation and our experimental results show that the performance of the wildfire simulation has been improved by 45% when compared with a static partitioning algorithm. We also evaluate the performance of ARM using two partitioning strategies: natural regions (NR) approach and a graph partitioning approach.