Despite of their successes, the results of first-principles quantum mechanical calculations contain inherent numerical errors that are caused by inadequate treatment of electron correlation, incompleteness of basis sets, relativistic effects or approximated exchange-correlation functionals. In this work, we develop a combined density-functional theory and neural-network correction (DFT-NEURON) approach to reduce drastically these errors, and apply the resulting approach to determine the standard Gibbs energy of formation ΔG ⁰ at 298 K for small- and medium-sized organic molecules. The root mean square deviation of the calculated ΔG ⁰ for 180 molecules is reduced from 22.3 kcal · mol^-1 to 3.0 kcal · mol^-1 for B3LYP/6-311+G(d,p). We examine further the selection of physical descriptors for the neural network.

Simulations of nanoelectronic devices with nonequilibrium Green's function are computationally very intensive, in particular, when combined with multiband approaches, such as the k·p methods. To reduce the cost and make the simulation of realistic devices tractable, we have developed a model order reduction method for the simulation of hole transport in silicon nanowires using three- and six-band k·p models. It is shown in this paper that, with a spurious band elimination process, the method can be readily extended to the eight-band case that enables us to simulate band-to-band tunneling devices. The method is demonstrated via constructing reduced models for indium arsenide (InAs) nanowires and simulation of I-V characteristics of InAs tunneling field-effect transistors (TFETs). The results indicate that significant model reduction can be achieved with good accuracy retained. The method is then applied to study InAs TFETs with different channel orientations and source-pocket TFETs with n-p-i-p doping profiles.

As an important part of the industrialization process, fully automated instrument monitoring and identification are experiencing an increasingly wide range of applications in industrial production, autonomous driving, and medical experimentation. However, digital instruments usually have multi-digit features, meaning that the numeric information on the screen is usually a multi-digit number greater than 10. Therefore, the accuracy of recognition with traditional algorithms such as threshold segmentation and template matching is low, and thus instrument monitoring still relies heavily on human labor at present. However, manual monitoring is costly and not suitable for risky experimental environments such as those involving radiation and contamination. The development of deep neural networks has opened up new possibilities for fully automated instrument monitoring; however, neural networks generally require large training datasets, costly data collection, and annotation. To solve the above problems, this paper proposes a new instrument monitoring method based on few-shot learning (FLIMM). FLIMM improves the average accuracy (ACC) of the model to 99% with only 16 original images via effective data augmentation method. Meanwhile, due to the controllability of simulated image generation, FLIMM can automatically generate annotation information for simulated numbers, which greatly reduces the cost of data collection and annotation.

In this paper we present a new multi-scale simulation scheme for next-generation electronic design automation for nano-electronics. The scheme features a combination of the first-principles quantum mechanical calculation, semi-classical semiconductor device simulation, compact model generation and circuit simulation. To demonstrate the feasibility of the proposed scheme, we apply our newly developed quantum mechanics/electromagnetics method to simulate the junctionless transistors. The simulation results are consistent with the experimental measurements and provide new insights on the depletion effect of the hetero-doped gate on the drain current. Based on the calculated I-V curves, a compact model is then constructed for the junctionless transistors. The validity of the compact model is further verified by the transient circuit simulation of an inverter.