Search

Large AI models, or foundation models, are models recently emerging with massive scales both parameter-wise and data-wise, the magnitudes of which often reach beyond billions. Once pretrained, large AI models demonstrate impressive performance in various downstream tasks. A concrete example is the recent debut of ChatGPT, whose capability has compelled people's imagination about the far-reaching influence that large AI models can have and their potential to transform different domains of our life. In health informatics, the advent of large AI models has brought new paradigms for the design of methodologies. The scale of multimodality data in the biomedical and health domain has been ever-expanding especially since the community embraced the era of deep learning, which provides the ground to develop, validate, and advance large AI models for breakthroughs in health-related areas. This article presents an up-to-date comprehensive review of large AI models, from background to their applications. We identify seven key sectors that large AI models are applicable and might have substantial influence, including 1) molecular biology and drug discovery; 2) medical diagnosis and decision-making; 3) medical imaging and vision; 4) medical informatics; 5) medical education; 6) public health; and 7) medical robotics. We examine their challenges in health informatics, followed by a critical discussion about potential future directions and pitfalls of large AI models in transforming the field of health informatics.

Inorganic trace minerals may exacerbate lipid peroxidation, thereby impacting lipid metabolism. This study aimed to compare the effects of inorganic and coated trace minerals in diets with different fat sources, on the performance, slaughter characteristics, and antioxidant status of broiler chickens. A total of 576 21-day-old Abor Acres broiler birds were randomly divided into four dietary treatment groups in a 2 (non-coated and coated trace minerals)×2 (soybean oil and lard) factorial design. Each treatment was replicated 12 times (12 birds per replicate). The results showed that coated minerals significantly improved the average daily gain (ADG) in weight and the feed conversion ratio (P

Cobalt species as active sites for photocatalytic reduction of CO$_{2}$ to valuable products such as methanol have received increasing attention, however, it remains a huge challenge to achieve the high activity. Herein, a pyrolysis-induced-vaporization strategy was successfully employed to fabricate Co/g-C$_{3}$N$_{4}$ single-atom catalysts (Co/g-C$_{3}$N$_{4}$ SACs) with surface Co atom loading up to 24.6 wt%. Systematic investigation of Co/g-C$_{3}$N$_{4}$ SACs formation process disclosed that concentrated-H$_{2}$SO$_{4}$ exfoliation of g-C$_{3}$N$_{4}$ nanosheets (g-C$_{3}$N$_{4}$ NSs) as the substrate followed by a two-step calcination process is essential to achieve ultrahigh metal loading. It was found that the ultrahigh-density of Co single-atom sites were anchored on the g-C$_{3}$N$_{4}$ substrate surface and coordinated with two nitrogen and one carbon atoms (Co-N$_{2}$C). These single dispersed Co-N$_{2}$C sites on the g-C$_{3}$N$_{4}$ surface were found to act not only as electron gathering centers but also as the sites of CO$_{2}$ adsorption and activation, subsequently, boosting the photocatalytic methanol generation during light irradiation. As a result, the methanol formation rate at 4 h (941.9 μmol g$^{-1}$) over Co/g-C$_{3}$N$_{4}$-0.2 SAC with 24.6 wt% surface Co loading was13.4 and 2.2 times higher than those of g-C$_{3}$N$_{4}$ (17.7 μmol g$^{-1}$) and aggregated CoOx/g-C$_{3}$N$_{4}$-0.2 (423.9 μmol g$^{-1}$), respectively. Simultaneously, H$_{2}$ (18.9 μmol g$^{-1}$ h$^{-1}$), CO (2.9 μmol g$^{-1}$ h$^{-1}$), CH$_{4}$ (3.4 μmol g$^{-1}$ h$^{-1}$), C$_{2}$H$_{4}$ (1.1 μmol g$^{-1}$ h$^{-1}$), C$_{3}$H$_{6}$ (1.4 μmol g$^{-1}$ h$^{-1}$), and CH$_{3}$OCH$_{3}$ (3.3 μmol g$^{-1}$ h$^{-1}$) products were detected over Co/g-C$_{3}$N$_{4}$-0.2 SAC. Besides, the photocatalytic activity of the Co/g-C$_{3}$N$_{4}$-0.2 SAC for the reduction of CO$_{2}$ to methanol was stable within 12-cycle experiments (~48 h). This work paves a strategy to boost the photoreduction CO$_{2}$ activity via loading ultrahigh surface density single atomically dispersed cobalt active sites.

Additional file 1: Table. S1. Groups and the number of ITS in complete genome of Pseudomonas.

Additional file 2: Table. S2. Strains in the double-blind experiment.

Seismic events, among many other natural hazards, reduce due functionality and exacerbate vulnerability of in-service buildings. Accurate modeling and prediction of building's response subjected to earthquakes makes possible to evaluate building performance. To this end, we leverage the recent advances in deep learning and develop a physics-guided convolutional neural network (PhyCNN) framework for data-driven seismic response modeling and serviceability assessment of buildings. The proposed PhyCNN approach is capable of accurately predicting building's seismic response in a data-driven fashion without the need of a physics-based analytical/numerical model. The basic concept is to train a deep PhyCNN model based on available seismic input-output datasets (e.g., from simulation or sensing) and physics constraints. The trained PhyCNN can then used as a surrogate model for structural seismic response prediction. Available physics (e.g., the law of dynamics) can provide constraints to the network outputs, alleviate overfitting issues, reduce the need of big training datasets, and thus improve the robustness of the trained model for more reliable prediction. The trained surrogate model is then utilized for fragility analysis given certain limit state criteria (e.g., the serviceability state). In addition, an unsupervised learning algorithm based on K-means clustering is also proposed to partition the limited number of datasets to training, validation and prediction categories, so as to maximize the use of limited datasets. The performance of the proposed approach is demonstrated through three case studies including both numerical and experimental examples. Convincing results illustrate that the proposed PhyCNN paradigm outperforms conventional pure data-based neural networks., 24 pages, 17 figures

To achieve closed trajectory motion planning of redundant manipulators, each joint angle has to be returned to its initial position. Most of the repeatable motion schemes have been proposed to solve kinematic problems considering only the initial desired position of each joint at first. Actually, it is very difficult for various joint angles of the robot arms to be positioned in the expected trajectory before moving. To construct an effective kinematic model, a novel optimal programming index based on a recurrent neural network is designed and analyzed in this paper. The repetitiveness and timeliness are presented and analyzed. Combining the kinematic equation constraint of manipulators, a repeatable motion scheme is formulated. In addition, the Lagrange multiplier theorem is introduced to prove that such a repeatable motion scheme can be converted into a time-varying linear equation. A finite-time neural network solver is constructed for the solution of the motion scheme. Simulation results for two different trajectories illustrate the accuracy and timeliness of the proposed motion scheme. Finally, two different repetitive schemes are compared and verified the optimal time for the novelty of the proposed kinematic scheme.

Terrorist attacks have become a growing threat worldwide in recent years. Explosive devices, the weapon of choice for a majority of terrorist attacks, significantly threaten civilian and military personnel. Accordingly it is very important to protect critical buildings against blast loads with the main goal of preventing loss of life of the occupants. The research detailed within this dissertation will investigate innovative smart structures, including the mitigation of damage and loss of life under blast loading through base isolated structures combined with supplemental passive control devices without compromising the innate seismic protection that base isolation provides. The focus of this dissertation is the development and simulation of multiple control strategies for multi-story structures subjected to surface blasts and seismic excitations. The goal is to study and improve the response of base isolated structures under blast loadings and simultaneously keep the same level or better performance under earthquakes through alternative energy dissipation systems.

Since social media gains ever increasing popularity, Online Social Networks have become important repositories for information retrieval. The concept of social search, therefore, is gradually being recognized as the next breakthrough in this field, and it is expected to dominate topics in industry. However, retrieving information from OSNs with high Quality of Experience is non-trivial as a result of the prevalence of mobile applications for social networking services. For the sake of shortening user perceived latency Web caching was introduced and has been studied extensively for years. Nevertheless, the previous works seldom focus on the Web caching solutions for social search. In the context of this master’s thesis project, emphasis is given to the design of a Web caching system which is used to cache public data from social media with the objective of improving the user experience in terms of the freshness of data and the perceived service latency. To be more specific, a Web caching strategy named Staleness Bounded LRU algorithm is proposed to limit the term of validity of the cached data. In addition, a Two-Level Web Caching System that adopts the SB-LRU algorithm is proposed in order for shortening the user perceived latency. Results of trace-driven simulations and performance evaluations demonstrate that serving clients with stale data is avoided and the user perceived latencies are significantly shortened when the proposed Web caching system is used in the use case of unauthenticated social search. Besides, the design idea in this project is believed to be helpful to the design of a Web caching system for social search, which is capable of caching user specific data for different clients.