We present Magnol.AI Copilot, an extension of the Magnol.AI digital biomarker platform that integrates multimodal large language models (LLMs) to transform digital health technology (DHT) trial dashboards into conversational systems. Copilot augments the platform with a multi-agent orchestration layer and vision-enabled LLMs that interpret visualizations, tabular summaries, and textual metadata. The system enables natural language queries and automatic generation of contextual insights, allowing researchers to interact with wearable data through dialogue rather than static inspection. A case study with an actigraphy device demonstrates Copilot’s ability to identify nightly compliance gaps and provide contextual explanations, reducing cognitive load compared to manual dashboard review. This work presents a novel integration of IoMT infrastructure with multimodal LLMs, advancing digital biomarker research toward conversational and accessible DHT trial platforms.

Over the past decade, the integration of digital technologies into clinical trials has fundamentally transformed the ability to monitor patients in real-time through wearable sensors, enabling the rapid capture of clinically relevant signals. This transformation has accelerated the development of medicines by providing faster, more objective measurements of therapeutic effects. Central to this technological shift is the Internet of Medical Things (IoMT) platform, with Magnol.Ai serving as a prime example. Magnol.Ai excels in continuously processing and converting vast amounts of digital data, including wearable sensor signals and electronic Patient-Reported Outcomes (ePROs), into actionable clinical insights for digital biomarker (dBM) research. Beyond its pivotal role in clinical research, Magnol.Ai demonstrates versatility as a scalable, real-time edge computing platform in other domains, such as real-time gait computing. By deploying advanced sensors and algorithms for real-time speed of movement tracking, Magnol.Ai underscores its capacity to support a broad range of applications within IoMT.

Connected clinical trials enable real-time connection with consented patients, allowing the capture of clinically meaningful signals via wearable sensors to accelerate the development of medicines. At the core of the full-stack technologies needed to support such trials is our digital data platform (DDP) — to continuously ingest, visualize, process, and transform a large amount of digital data, including wearable sensor signals and electronic Patient-Reported Outcomes (ePRO)s into meaningful clinical measures. This paper presents the integrated systems and technologies we have developed to establish such a cloud-based platform. The main advantages of the DDP include 1) interactive exploration and navigation of large volumes of digital data at scale, 2) real-time data monitoring to promptly track data quality and compliance, and 3) novel data delivery methods to allow seamless access to large-scale raw sensor signals, processed signals, and aggregated digital biomarker measures for endpoints analysis. Finally, two use scenarios developed and deployed in the DDP will be demonstrated.

Large bibliographic data sets hold the promise of revolutionizing the scientific enterprise when combined with state-of-the-science computational capabilities. Providing high-quality data services for large network datasets such as the Microsoft Academic Graph, which contains more than two billion citation links, poses significant difficulties for universities. Data systems based on the property graph model are capable of delivering efficient graph query services for large networks. However, real-life queries often combine multiple types of data models. To satisfy the needs of different user groups, we developed and deployed a cloud-based data system consisting of scalable graph and text-indexed query engines. For non-expert users, the property graph model also presents a technological barrier. To alleviate the steep learning curve, we designed an intuitive graphical user interface for query-building. For advanced users, a scalable notebook service in our platform provides a more flexible computing environments where the query results can be further analyzed. These systems form the data-backbone of the Collaborative Archive and Data Research Environment (CADRE), which provides efficient and high-quality bibliographic data services to eleven large public universities in North America.

Over the past decade, cyberinfrastructure community like XSEDE has substantially fostered and enriched knowledge discovery of scholars, researchers, and engineers from a variety of domains through enabling access to advanced computing systems, where continuing support for classic packages and parallel computing frameworks (e.g., MPI and OpenMP) has been well established. However, with the rise of "Big Data" era, an ever increasing demand from user community is the desire to run sophisticated, state-of-the-art distributed frameworks that handle various data related tasks. Examples include Hadoop and Spark for data processing and analytics, Cassandra and Redis for scalable on-disk and in-memory data stores, Apache Airflow for distributed workflow engine, just to name a few. Though by design such frameworks provision high-level, user friendly programming APIs for business logic composition, their deployment process oftentimes is both complex and complicated, requiring expertise well beyond what the majority of cyberinfrastructure users may have. To bridge the gap, in this paper we propose the concept of provisioning such frameworks through "cyberinfrastructure managed", system wide services, especially through leveraging "high level, complexity hiding interfaces" principle when designing interfaces that are exposed to users for framework setup and shutdown. In particular, we use Spark-as-a-Service at Indiana University as a concrete case study to illustrate how we applied the design principles. Furthermore, to demonstrate the generality of the design, we showcase Cassandra-as-a-Service, a work-in-progress prototype.

Geometric problems of interest to mathematical visualization applications involve changing structures, such as the moves that transform one knot into an equivalent knot. In this paper, we describe mathematical entities (curves and surfaces) as link-node graphs, and make use of energy-driven relaxation algorithms to optimize their geometric shapes by moving knots and surfaces to their simplified equivalence. Furthermore, we design and conFigure parallel functional units in the relaxation algorithms to accelerate the computation these mathematical deformations require. Results show that we can achieve significant performance optimization via the proposed threading model and level of parallelization.

MapReduce is a parallel framework which has been widely adopted for conducting large-scale data analytics. In cases where analysis of multiple millions of books must be analyzed using federally funded high performance computing (HPC) resources, the framework fails to port directly. We propose a solution that builds off of MapReduce for use on a HPC system that preserves the key-value semantics of map-reduce while supporting the random access of query access for subsetting Big Data datasets, and at same time hosting the service using the storage medium found in HPC architectures (parallel file systems) for reduced latencies. Experimental results demonstrate Horme's good performance in the HPC setting, with up to 41.4% faster than NoSQL based solution in random access scenario.

Dentistry is a particularly complex and sophisticated applied science; many problems have to be solved by analyzing intensive longitudinal data. For example, dynamic carious lesion assessment requires dental researchers to perform knowledge discovery in a situation with multiple specimens across different experimental phases. The technological development and availability of cyberinfrastructure today can enable dental researchers to perform existing procedures far faster and more accurately than ever. This paper uses dynamic carious lesion activity assessment as a case study, to illustrate how visual computing on advanced cyberinfrastructure can expand beyond statistical number crunching and information retrieval to make an imaginative and creative contribution to some aspects of dental science. Our work focuses on the generation of BIG pictures on cyberinfrastructure and the presentation of derived dental structures in an interactive means, which combine to allow researchers to navigate from observation to qualitative discovery and then to quantitative assessment with multiple variables and degrees of freedom. Our work has seen early use by our collaborators in oral health research, where our system has been used to pose and answer domain-specific questions for quantitative assessment of dynamic carious lesion activities.

Mining frequent subsequences of patterns, or sequential pattern mining, has wide application in customer shopping sequence analysis, web log stream analysis, multi-modal behavioral studies, to name a few. To detect unknown, anomalous, and unexpected patterns from large-scale interval-based temporal data without complete a priori knowledge is challenging. In this paper, we present a framework - PESMiner which allows parallel and quantitative mining of sequential patterns at scale. Whereas most existing sequential mining algorithms can only find sequential orders of temporal events, our work presents a novel interactive temporal data mining algorithm capable of extracting precise temporal properties of sequential patterns. Furthermore, our work provides a unified parallel solution that scales our algorithms to larger temporal data sets by exploiting iterative MapReduce tasks. Comprehensive performance evaluations demonstrate that PESMiner significantly outperforms existing interval-based mining algorithms in terms of both quality (i.e. accuracy, precision, and recall) and scalability.

In this paper, we present a mathematical visualization paradigm for exploring curves embedded in 3D and surfaces in 4D mathematical world. The basic problem is that, 3D figures of 4D mathematical entities often twist, turn, and fold back on themselves, leaving important properties behind the surface sheets. We propose an interactive system to visualize the topological features of the original 4D surface by slicing its 3D figure into a series of feature diagram. A novel 4D visualization interface is designed to allow users to control 4D topological shapes via the collection of diagram handles using the established curve manipulation mechanism. Our system can support rich mathematical interaction of 4D mathematical objects which is very difficult with any existing approach. We further demonstrate the effectiveness of the proposed visualization tool using various experimental results and cases studies.

Online social networking communities may exhibit highly complex and adaptive collective behaviors. Since emotions play such an important role in human decision making, how online networks modulate human collective mood states has become a matter of considerable interest. In spite of the increasing societal importance of online social networks, it is unknown whether assortative mixing of psychological states takes place in situations where social ties are mediated solely by online networking services in the absence of physical contact. Here, we show that the general happiness, or subjective well-being (SWB), of Twitter users, as measured from a 6-month record of their individual tweets, is indeed assortative across the Twitter social network. Our results imply that online social networks may be equally subject to the social mechanisms that cause assortative mixing in real social networks and that such assortative mixing takes place at the level of SWB. Given the increasing prevalence of online social networks, their propensity to connect users with similar levels of SWB may be an important factor in how positive and negative sentiments are maintained and spread through human society. Future research may focus on how event-specific mood states can propagate and influence user behavior in "real life."

Inspired by human immune system, a concentration based feature construction (CFC) approach which utilizes a two-element concentration vector as the feature vector is proposed for spam detection in this paper. In the CFC approach, dasiaselfpsila and dasianon-selfpsila concentrations are constructed by using dasiaselfpsila and dasianon-selfpsila gene libraries, respectively, and subsequently are used to form a vector with two elements of concentrations for characterizing the e-mail efficiently. As a result, the design of classifier actually amounts to establishing a mapping between two real-value inputs and one binary output. The classification of the e-mail is considered as an optimization problem aiming at minimizing a formulated cost function. A clonal particle swarm optimization (CPSO) algorithm proposed by the leading author is also employed for this purpose. Several classifiers including linear discriminant, multi-layer neural networks and support vector machine are used to verify the effectiveness and robustness of the CFC approach. Experimental results demonstrate that the proposed CFC approach not only has a very much fast speed but also gives 97% and 99% of accuracy just using a two-element concentration feature vector on corpus PU1 and Ling, respectively.