Search

Feature learning plays an important role in many machine learning tasks. As a common implementation for feature learning, the auto-encoder has shown excellent performance. However it also faces several challenges among which a notable one is how to reduce its generalization error. Different approaches have been proposed to solve this problem and Bagging is lauded as a possible one since it is easily implemented while can also expect outstanding performance. This paper studies the problem of integrating different prediction models by bagging auto-encoder-based classifiers in order to reduce generalization error and improve prediction performance. Furthermore, experimental study on different datasets from different domains is conducted. Several integration schemas are empirically evaluated to analyse their pros and cons. It is believed that this work will offer researchers in this field insight in bagging auto-encoder-based classifiers.

Two-dimensional transition metal dichalcogenides (TMDs) are promising low-dimensional materials which can produce diverse electronic properties and band alignment in van der Waals heterostructures. Systematic density functional theory (DFT) calculations are performed for 24 different TMD monolayers and their bilayer heterostacks. DFT calculations show that monolayer TMDs can behave as semiconducting, metallic or semimetallic depending on their structures; we also calculated the band alignment of the TMDs to predict their alignment in van der Waals heterostacks. We have applied the charge equilibration model (CEM) to obtain a quantitative formula predicting the highest occupied state of any type of bilayer TMD heterostacks (552 pairs for 24 TMDs). The CEM predicted values agree quite well with the selected DFT simulation results. The quantitative prediction of the band alignment in the TMD heterostructures can provide an insightful guidance to the development of TMD-based devices.

The control of domain morphology and defect level of synthesized transition metal dichalcogenides (TMDs) is of crucial importance for their device applications. However, current TMDs synthesis by chemical vapor deposition and molecular beam epitaxy is in an early stage of development, where much of the understanding of the process-property relationships is highly empirical. In this work, we use a kinetic Monte Carlo coupled with first principles calculations to study one specific case of the deposition of monolayer WSe2 on graphene, which can be expanded to the entire TMD family. Monolayer WSe2 domains are investigated as a function of incident flux, temperature and precursor ratio. The quality of the grown WSe2 domains is analyzed by the stoichiometry and defect density. A phase diagram of domain morphology is developed in the space of flux and the precursor stoichiometry, in which the triangular compact, fractal and dendritic domains are identified. The phase diagram has inspired a new synthesis strategy for large TMD domains with improved quality.

Tungsten diselenide (WSe2) is a semiconducting, two-dimensional (2D) material that has gained interest in the device community recently due to its electronic properties. The synthesis of atomically thin WSe2, however, is still in its infancy. In this work we elucidate the requirements for large selenium/tungsten precursor ratios and explain the effect of nucleation temperature on the synthesis of WSe2 via metal-organic chemical vapor deposition (MOCVD). The introduction of a nucleation-step prior to growth demonstrates that increasing nucleation temperature leads to a transition from a Volmer–Weber to Frank–van der Merwe growth mode. Additionally, the nucleation step prior to growth leads to an improvement of WSe2 layer coverage on the substrate. Finally, we note that the development of this two-step technique may allow for improved control and quality of 2D layers grown via CVD and MOCVD processes.

Two novel iridium(III) complexes, [Ir(dfppy)(2)(pmc)] and [Ir(ppy)(2)(pmc)] (dfppy = 2-(4',6'-difluoro-phenyl)pyridine, ppy = 1-phenyl-pyridine), were designed and synthesized using 2-carboxyl-pyrimidine (Hpmc) as an ancillary ligand. Single crystals were obtained and characterized by single crystal X-ray diffraction. The tetrametallic complexes {[(C^N)(2)Ir(μ-pmc)](3)EuCl(3)} (C^N = dfppy, ppy) were synthesized using the iridium(III) complexes as "ligands". Photophysical and theoretical studies indicate that [Ir(dfppy)(2)(pmc)] is more suitable for sensitizing the emission of Eu(III) ions than [Ir(ppy)(2)(pmc)].