Deep Transfer Learning for the Multilabel Classification of Chest X-ray Images.

Chest X-ray (CXR) is widely used to diagnose conditions affecting the chest, its contents, and its nearby structures. In this study, we used a private data set containing 1630 CXR images with disease labels; most of the images were disease-free, but the others contained multiple sites of abnormalities. Here, we used deep convolutional neural network (CNN) models to extract feature representations and to identify possible diseases in these images. We also used transfer learning combined with large open-source image data sets to resolve the problems of insufficient training data and optimize the classification model. The effects of different approaches of reusing pretrained weights (model finetuning and layer transfer), source data sets of different sizes and similarity levels to the target data (ImageNet, ChestX-ray, and CheXpert), methods integrating source data sets into transfer learning (initiating, concatenating, and co-training), and backbone CNN models (ResNet50 and DenseNet121) on transfer learning were also assessed. The results demonstrated that transfer learning applied with the model finetuning approach typically afforded better prediction models. When only one source data set was adopted, ChestX-ray performed better than CheXpert; however, after ImageNet initials were attached, CheXpert performed better. ResNet50 performed better in initiating transfer learning, whereas DenseNet121 performed better in concatenating and co-training transfer learning. Transfer learning with multiple source data sets was preferable to that with a source data set. Overall, transfer learning can further enhance prediction capabilities and reduce computing costs for CXR images.

Molecular aluminum hydrides identified by inelastic neutron scattering during H2 regeneration of catalyst-doped NaAlH4.

Catalyst-doped sodium aluminum hydrides have been intensively studied as solid hydrogen carriers for onboard proton-exchange membrane (PEM) fuel cells. Although the importance of catalyst choice in enhancing kinetics for both hydrogen uptake and release of this hydride material has long been recognized, the nature of the active species and the mechanism of catalytic action are unclear. We have shown by inelastic neutron scattering (INS) spectroscopy that a volatile molecular aluminum hydride is formed during the early stage of H2 regeneration of a depleted, catalyst-doped sodium aluminum hydride. Computational modeling of the INS spectra suggested the formation of AlH3 and oligomers (AlH3)n (Al2H6, Al3H9, and Al4H12 clusters), which are pertinent to the mechanism of hydrogen storage. This paper demonstrates, for the first time, the existence of these volatile species.

Engineering Pt in ceria for a maximum metal-support interaction in catalysis.

Conventional supported metal catalysts are metal nanoparticles deposited on high surface area oxide supports with a poorly defined metal-support interface. Typically, the traditionally prepared Pt/ceria catalyzes both methanation (H2/CO to CH4) and water-gas shift (CO/H2O to CO2/H2) reactions. By using simple nanochemistry techniques, we show for the first time that Pt or PtAu metal can be created inside each CeO2 particle with tailored dimensions. The encapsulated metal is shown to interact with the thin CeO2 overlayer in each single particle in an optimum geometry to create a unique interface, giving high activity and excellent selectivity for the water-gas shift reaction, but is totally inert for methanation. Thus, this work clearly demonstrates the significance of nanoengineering of a single catalyst particle by a bottom-up construction approach in modern catalyst design which could enable exploitation of catalyst site differentiation, leading to new catalytic properties.