Search

There have been impressive advances in the physical and mathematical modelling of complex physiological systems in the last few decades, with the potential to revolutionise personalised healthcare with patient-specific evidence-based diagnosis, risk assessment and treatment decision support using digital twins. However, practical progress and genuine clinical impact hinge on successful model calibration, parameter estimation and uncertainty quantification, which calls for novel innovative adaptions and methodological extensions of contemporary state-of-the-art inference techniques from Statistics and Machine Learning. In the present study, we focus on two computational fluid-dynamics (CFD) models of the blood systemic and pulmonary circulation. We discuss state-of-the-art emulation techniques based on deep learning and Gaussian processes, which are coupled with established inference techniques based on greedy optimisation, simulated annealing, Markov Chain Monte Carlo, History Matching and rejection sampling for computationally fast inference of unknown parameters of the CFD models from blood flow and pressure data. The inference task was set as a competitive challenge which the participants had to conduct within a limited time frame representative of clinical requirements. The performance of the methods was assessed independently and objectively by the challenge organisers, based on a ground truth that was unknown to the method developers. Our results indicate that for the systemic challenge, in which an idealised case of noise-free data was considered, the relative deviation from the ground-truth in parameter space ranges from 10−5% (highest-performing method) to 3% (lowest-performing method). For the pulmonary challenge, for which noisy data was generated, the performance ranges from 0.9% to 7% deviation for the parameter posterior mean, and from 35% to 570% deviation for the parameter posterior variance.

Non-intrusive load monitoring （NILM） not only makes the flow of electric energy transparent but also simplifies the installation process of smart meters， effectively reducing the cost of load monitoring. To enhance the accuracy of load recognition in NILM， a method for load recognition based on data augmentation and threshold-free recurrence plot （RP） is proposed. a denoising diffusion probability model （DDPM） is utilized to augment the load data of small samples to enhance the robustness of the load recognition method. Furthermore， a threshold-free RP， achieved by removing the Heaviside function of the recurrence graph， efficiently represents load characteristics. This is combined with a Transformer deep learning network to construct a load recognition framework. The proposed method is applied to three real-world datasets， and experimental results demonstrate its effectiveness in improving load recognition accuracy and enhancing classification performance.

Achieving subwavelength optical focusing is of great importance in nanophotonics. However, achieving focusing with both a small focal spot size and a large focal length remains elusive so far. Here, a large focal length planar focusing device is presented, utilizing highly oriented Dyakonov polaritons in hyperbolic metamaterial with periodic silver rings as the excitation source. Experimental results show that by controlling the size of the excitation sources, the focal length can reach 6 λ 0 (where λ 0 is the illumination wavelength), and the focal spot size can be reduced to λ 0 /10. This method provides new prospects for planar polariton optical applications. [ABSTRACT FROM AUTHOR]

Based on the food equivalent unit (FEU), this article analyzed Chinese food consumption patterns, spatial mismatch, and production potential to explore agricultural reform strategies. Assessing production–demand mismatch involved the spatial mismatch model, drawing data from statistical yearbooks. Calculations of food production potential utilized the CASA model and the Thornthwaite Memorial model, with net primary productivity (NPP) derived from remote sensing data as indicators. The results showed that livestock product consumption is on the rise, and the spatial mismatch index for herbivorous livestock products was the largest, ranging from 22.81 to 98.12 in 2019. The mismatched degree distribution of rations and food-consuming livestock products showed a trend of increasing on both sides, with the Hu Huanyong line as the center line. Production factors played a predominant role in food production-to-demand mismatch. Climatic productivity and actual productivity decreased from the southeast to northwest in space in 2019, and human activities significantly impacted productivity. When grassland agriculture is pursued as the adjustment orientation, the production potential can reach up to 4540.76 × 107 kg·FEU. Moreover, a grassland agriculture plan was devised, prioritizing its development in the developed southern regions. [ABSTRACT FROM AUTHOR]