When modeling related tasks in computer vision, Multi-Task Learning (MTL) can outperform Single-Task (STL) due to its ability capture intrinsic relatedness among tasks. However, MTL may encounter the insufficient training problem, i.e., some non-optimal situation compared with STL. A series of studies point out that too much gradient noise would lead performance degradation STL, however, scenar...

Molecular dynamics simulation has been carried out to explore the configuration and quantity threshold of multiple graphene nanoribbons (GNRs) in single-walled carbon nanotube (SWCNT). The simulation results showed that several GNRs tangled together to form a perfect spiral structure to maximize the π-π stacking area when filling inside SWCNT. The formation of multiple helical configuration is ...

Transport properties of sub-5 nm-wide graphene nanoribbons (GNRs) are investigated by using atomistic non-equilibrium Green’s function (NEGF) simulations and semiclassical mobility simulations of large ensembles of randomly generated nanoribbons. Realistic GNRs with dimensions targeting the 12 nm CMOS node are investigated by accounting for edge defects, vacancies and potential fluctuations. Ef...

Vibrational dephasing of the lowest energy electronic excitations in the perfect (16,16) graphene nanoribbon (GNR) and those with the C2-bond insertion and rotation defects is studied with ab initio molecular dynamics. Compared to single-walled carbon nanotubes (SWCNTs) of similar size, GNRs shows very different properties. The dephasing in the ideal GNR occurs twice faster than that in the SWC...

2D penta-graphene sheets were cut along typical crystallographic orientations in order to construct various penta-graphene nanoribbon (P-GNR) models, and their electronic structures and magnetic properties were systemically investigated. It was demonstrated that P-GNRs are very versatile with rich and unique electronic and magnetic properties. In particular, bipolar magnetic semiconducting feat...

We investigate high-field transport in graphene nanoribbons (GNRs) on SiO(2), up to breakdown. The maximum current density is limited by self-heating, but can reach >3 mA/μm for GNRs ~15 nm wide. Comparison with larger, micron-sized graphene devices reveals that narrow GNRs benefit from 3D heat spreading into the SiO(2), which enables their higher current density. GNRs also benefit from lateral...

Bottom-up assembled nanomaterials and nanostructures allow for the studies of rich unprecedented quantum-related mesoscopic transport phenomena. However, it can be difficult to quantify correlations between geometrical or structural parameters obtained from advanced microscopy measured electrical characteristics when they are made into macroscopic devices. Here, we propose a strategy connect na...

Graphene oxide nanoribbons (GONR) with regular edges have been prepared by a facile hydrothermal method. The relatively high yield and convenient preparation of GONR eventually make graphene nanoribbons (GNR) accessible in the field of composite materials where bulk quantities of nanoribbons are required.

Simonov, K. 2016. Effect of Substrate on Bottom-Up Fabrication and Electronic Properties of Graphene Nanoribbons. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1385. 101 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-554-9610-4. Taking into account the technological demand for the controlled preparation of atomically precise graphe...