Background and Obejctives: Multi-task learning is a widespread mechanism to improve the learning of multiple objectives with a shared representation in one deep neural network. In multi-task learning, it is critical to determine how to combine the tasks loss functions. The straightforward way is to optimize the weighted linear sum of multiple objectives with equal weights. Despite some studies that have attempted to solve the realtime multi-person pose estimation problem from a 2D image, major challenges still remain unresolved. Methods: The prevailing solutions are two-stream, learning two tasks simultaneously. They intrinsically use a multi-task learning approach for predicting the confidence maps of body parts and the part affinity fields to associate the parts to each other. They optimize the average of the two tasks loss functions, while the two tasks have different levels of difficulty and uncertainty. In this work, we overcome this problem by applying a multi-task objective that captures task-based uncertainties without any additional parameters. Since the estimated poses can be more certain, the proposed method is called “CertainPose”. Results: Experiments are carried out on the COCO keypoints data sets. The results show that capturing the task-dependent uncertainty makes the training procedure faster and causes some improvements in human pose estimation. Conclusion: The highlight advantage of our method is improving the realtime multi-person pose estimation without increasing computational complexity.

Distillation columns are essential in the chemical and process industries for separating liquid mixtures based on differences in the volatility of their components. Achieving the desired product compositions requires effective modelling and control strategies. This work focuses on developing a dynamic mathematical model for the UOP3CC Armfield binary distillation column used in the laboratory. The model incorporates mass and energy balance equations and considers key factors such as heat transfer in the condenser and reboiler, tray hydraulics, and vapor-liquid equilibrium. The model parameters are estimated and validated using real-time experimental data. The control strategy aims to regulate the temperature profiles to maintain the desired ethanol composition in the top product. This work explores the implementation of Proportional-Integral (PI), Model Predictive Control (MPC), and Robust setpoint Tracking Disturbance rejection with Aggressiveness (RTD-A) controllers to optimize process performance. Model order reduction techniques simplify the complex process dynamics, while a decoupler is used to manage interactions between control loops. The performance of the PI, MPC, and RTD-A controllers is compared using metrics like setpoint tracking, disturbance rejection, and computational efficiency. Experiments are conducted to establish the relationship between top tray temperature and distillate purity, leading to the identification of a temperature setpoint that maximizes ethanol purity. This setpoint is validated in a closed-loop system, with the physical model achieving a steady-state error of just 0.19%. The results emphasize the significance of accurate modelling in improving the performance of the UOP3CC distillation column.

Aim: Efforts to explore biomarkers and biological pathways involved in the disease are needed to improve colorectal cancer (CRC) diagnosis and alternative treatments Background: The fourth common malignancy in the world is colorectal cancer. The over-all burden is predicted to rise by 2030. Methods: In the current study, nine genes were selected. Previously, a panel of genes by Agendia, a classifier of robust gene expression (ColoPrint), was determined to significantly improve the prognostic accuracy of pathologic factors in stage II and III colorectal cancer patients. Five genes, including Ppara, Mctp1, Pyroxd1, Il2r, and Cyfip2, from this panel and four other genes which were not in this panel but were cited abundantly in the literature were selected. Then, expression levels of the selected genes in CRC tissue were compared with levels in adjacent normal tissue. To identify the pathways involved in CRC, gene set enrichment analysis was subsequently performed. Furthermore, to illustrate the relationship between genes in this disease, the cross-shaped co-expression pattern and gene regulatory network were determined using computational methods. Results: This research found that the pairs of genes: {Il2r, Cyfip2}, {Foxm1, Ppara}, {Mctp1, Ctsc} and {Pyroxd1, Cyfip2} are functionally related. Furthermore, two differentially expressed gene pairs {Foxm1, Ppara} and {Il2r, Cyfip2} are involved in the vascular endothelial growth factor receptor signaling pathway and the purine ribonucleoside diphosphate metabolic process, respectively. Conclusion: This research found that the combination of computational analysis and laboratory data provided the opportunity to better characterize the relation between central colorectal cancer genes as well as possible pathways involved in the colorectal cancer.