Daphne Teck Ching Lai

This research aims to study and assess state-of-the-art physics-informed neural networks (PINNs) from different researchers’ perspectives. The PRISMA framework was used for a systematic literature review, and 120 research articles from the computational sciences and engineering domain were specifically classified through a well-defined keyword search in Scopus and Web of Science databases. Through bibliometric analyses, we have identified journal sources with the most publications, authors with high citations, and countries with many publications on PINNs. Some newly improved techniques developed to enhance PINN performance and reduce high training costs and slowness, among other limitations, have been highlighted. Different approaches have been introduced to overcome the limitations of PINNs. In this review, we categorized the newly proposed PINN methods into Extended PINNs, Hybrid PINNs, and Minimized Loss techniques. Various potential future research directions are outlined based on the limitations of the proposed solutions.

Peatlands are important carbon sinks, therefore their degradation mainly due to lowering of the water table, has an adverse effect to the carbon cycle and largely burden the atmosphere. Annually, extensive areas of these peatlands are affected by wildfires, therefore large peatland areas have been degraded, directly or indirectly related to anthropogenic activities, largely contributing to greenhouse gas emissions. Some of the most pristine tropical peatlands worldwide occur primarily at the coastal lowlands of Southeast Asia. In terms of their geological features and evolution, these sites are generally understudied despite covering more than half of the global area of tropical peatlands. This review compiles geological data from 52 peatlands from insular and continental Southeast Asia, providing a comprehensive geological dataset for future research. The Southeast Asian peatlands are mostly ombrogenous and hence poorly supplied by nutrients. During the Last Glacial Period (ca. 115,000–11,000 years ago), extensive areas were exposed because of the lowering of the seawater level, which caused a consequent lowering of the groundwater table landwards; the forests were under severe threat, mainly due to dry conditions, thus resulting in the retreat of the forest boundaries inland. This was an unfavourable environment for peatland formation and hence, most of the Southeast Asian peatlands were formed right after that period. Almost 40% of the reviewed sites are located on Borneo Island, highlighting the importance of Bornean peatlands, where many peatlands are already deforested and drained and converted to plantations. Overall, the available geological data from the Southeast Asian peatlands is incomplete and non-comparable to each other because each study has a different focus. Details, such as the type of peat-forming plants, age of peat, peat thickness, substrate type and the pH value are not reported systematically in approximately 30% of the reviewed sites, while other important geological data, such as the ash yield and the carbon content of peat are only reported in 30% and 10% of the reviewed sites, respectively. Characterisation of peatlands using data imputation and principal component analysis (PCA) is based on three physical parameters (maximum ash yield, maximum thickness and oldest age), and includes the study of their relation in terms of climatic periods, peatland type, region and substrate. It is observed that peatlands which were developed in warming periods share similar physical parameters (such as accumulation rates, ash yield, peatland type, and environment of the substrates). With better data reporting on these parameters, the PCA analysis can provide a more accurate reflection of peatland characteristics and their relationships. The study aims to raise awareness on the importance and vulnerability of the Southeast Asian peatlands and to highlight their role in the global climate fluctuations.

Abstract Thermal maturity, organic richness and kerogen typing are very important parameters to be evaluated for source rock characterization. Due to the difficulties of high cost geochemical analyses and the unavailability of rock samples, it was necessary to examine and test many different method and techniques to help in the prediction of TOC values as well as other maturity indicators in case of missing or absence of geochemical data. Integrated study of machine learning techniques and well-log data has been applied on Cretaceous–Paleocene formations in the Taranaki Basin, New Zealand. A novel approach of maturity prediction using T max and vitrinite reflectance (VR%) is the first and preliminary objective of this research. Moreover, the organic richness or the total organic carbon (TOC) content has been predicted as well. Geochemical and well-log data collected from the Cretaceous Rakopi and North Cape formations and Paleocene Mangahewa Formation have been processed and prepared to apply the machine learning techniques. Five machine learning techniques, namely Bayesian regularization for feed-forward neural networks (BRNNs), random forest (RF), support vector machine (SVM) for regression, linear regression (LR) and Gaussian process regression (GPR), were employed for prediction of TOC, T max and VR, and their results have been compared. For TOC prediction, the best model achieved the coefficient of determination ( R 2 ) value of 0.964 using RF model. For T max prediction, BRNN with one hidden layer achieved the R 2 value of 0.828. BRNN with two hidden layers produced the best model for VR prediction achieving R 2 = 0.636. A comparison of five ML techniques showed that all of these techniques performed exceedingly well for TOC prediction with a value of R 2 > 0.96. In contrast, BRNN with one hidden layer was the only ML technique able to achieve R 2 > 0.8 for T max and BRNN with two hidden layers was the only ML technique able to achieve R 2 > 0.6 for VR prediction. Therefore, this research provides a strong empirical evidence that ML techniques can capture the nonlinear relationship between the well-log data and TOC as well as the maturity indicators which may not be fully understood by existing linear models.

BACKGROUND: Edge-to-edge approximation of the mitral valve leaflets (Alfieri procedure) is a novel surgical treatment for patients with ischemic mitral regurgitation (IMR). Long-term durability may be limited if abnormal mitral leaflet stresses result from this procedure. The aim of the current study was to measure Alfieri stitch tension (F(A)) and to explore its geometric determinants in an ovine model of acute IMR as a reflection of the mitral leaflet stresses imposed by the procedure. METHODS AND RESULTS: Eight sheep were studied immediately after surgical placement of (1) a force transducer interposed between sutures approximating the central leaflet edges and (2) radiopaque markers around the mitral annulus and leaflet edges. Computer-aided analysis of videofluorograms was used to obtained 3D marker coordinates. Simultaneous measurements of F(A), septal-lateral annular dimension (L(S-L)), leaflet edge separation (L(SEP)), anterior (L(AL)) and posterior (L(PL)) leaflet length, and hemodynamic variables were obtained at baseline (CTL) and during acute IMR (circumflex artery occlusion). F(A) was significantly elevated throughout the cardiac cycle during IMR compared with CTL, with maximum F(A) in diastole (0.26+/-0.05 versus 0.46+/-0.08 N, CTL versus IMR; P<0.05). Multivariable analysis revealed L(S-L) as the single independent predictor of maximum F(A) (P<0.001). Positive linear correlations were shown between values of F(A) and L(AL) and L(PL) (dependent variables). CONCLUSIONS: These experimental data demonstrate higher F(A) during IMR and cyclic changes in F(A) closely paralleling changes in L(S-L), eg, being greatest in diastole when the annulus is largest. Increased F(A) during IMR is probably indicative of successful therapeutic intent, but higher diastolic leaflet stresses resulting from persistent or progressive mitral annular dilatation may adversely affect repair durability. This indirectly implies that concomitant mitral ring annuloplasty should be added to the Alfieri repair.