Transfer learning in deep neural network based under-sampled MR image reconstruction

In Magnetic Resonance Imaging (MRI), the success of deep learning-based under-sampled MR image reconstruction depends on: (i) size of the training dataset, (ii) generalization capabilities of the trained neural network. Whenever there is a mismatch between the training and testing data, there is a need to retrain the neural network from scratch with thousands of MR images obtained using the same protocol. This may not be possible in MRI as it is costly and time consuming to acquire data. In this research, a transfer learning approach i.e. end-to-end fine tuning is proposed for U-Net to address the data scarcity and generalization problems of deep learning-based MR image reconstruction. First the generalization capabilities of a pre-trained U-Net (initially trained on the human brain images of 1.5 T scanner) are assessed for: (a) MR images acquired from MRI scanners of different magnetic field strengths, (b) MR images of different anatomies and (c) MR images under-sampled by different acceleration factors. Later, end-to-end fine tuning of the pre-trained U-Net is proposed for the reconstruction of the above-mentioned MR images (i.e. (a), (b) and (c)). The results show successful reconstructions obtained from the proposed method as reflected by the Structural SIMilarity index, Root Mean Square Error, Peak Signal-to-Noise Ratio and central line profile of the reconstructed images.     

Distribution dependent stochastic differential equations

Due to their intrinsic link with nonlinear Fokker-Planck equations and many other applications, distribution dependent stochastic differential equations (DDSDEs) have been intensively investigated. In this paper, we summarize some recent progresses in the study of DDSDEs, which include the correspondence of weak solutions and nonlinear Fokker-Planck equations, the well-posedness, regularity estimates, exponential ergodicity, long time large deviations, and comparison theorems.     

Embedded exponential-type low-regularity integrators for KdV equation under rough data

BIT Numerical Mathematics - In this paper, we introduce a novel class of embedded exponential-type low-regularity integrators (ELRIs) for solving the KdV equation and establish their optimal...     

Event-triggered design for discrete-time nonlinear systems with control constraints

Nonlinear Dynamics - In order to solve the constrained-input problem and reduce the computing resources, a novel event-triggered optimal control method is proposed for a class of discrete-time...     

Effective suppression of beta oscillation in Parkinsonian state via a noisy direct delayed feedback control scheme

This work explores the function of the noisy direct delayed feedback(NDDF)control strategy in suppressing the pathological oscillations in the basal ganglia(BG)with Parkinson’s disease(PD).Deep brain stimulation(DBS)alleviates the PD state fantastically.However,due to its unclear mechanism and open-loop characteristic,it is challenging to further improve its effects with lower energy expenditure.The noise stimulus performs competitively in alleviating the PD state theoretically,but it cannot adapt to the neural condition timely and automatically due to its open-loop control scheme.The direct delayed feedback(DDF)control strategy is able to disturb excessive synchronous effectively.Therefore,the NDDF control strategy is proposed and researched based on a BG computational model,which can reflect the intrinsic properties of the BG neurons and their connections with thalamic neurons.Simulation results show that the NDDF control strategy with optimal parameters is effective in removing the pathological beta oscillations.By comparison,we find the NDDF control strategy performs more excellent than DDF in alleviating PD state.Additionally,we define the multiple-NDDF control strategy and find that the multiple-NDDF with appropriate parameters performs better than NDDF.The obtained results contribute to the cure for PD symptoms by optimizing the noise-induced improvement of the BG dysfunction.     

Shedding vortex simulation method based on viscous compensation technology research

Owing to the influence of the viscosity of the flow field, the strength of the shedding vortex decreases gradually in the process of backward propagation. Large-scale vortexes constantly break up, forming smaller vortexes. In engineering, when numerical simulation of vortex evolution process is carried out, a large grid is needed to be arranged in the area of outflow field far from the boundary layer in order to ensure the calculation efficiency. As a result, small scale vortexes at the far end of the flow field cannot be captured by the sparse grid in this region, resulting in the dissipation or even disappearance of vortexes. In this paper, the effect of grid scale is quantified and compared with the viscous effect through theoretical derivation. The theoretical relationship between the mesh viscosity and the original viscosity of the flow field is established, and the viscosity term in the turbulence model is modified. This method proves to be able to effectively improve the intensity of small-scale shedding vortexes at the far end of the flow field under the condition of sparse grid. The error between the simulation results and the results obtained by using fine mesh is greatly reduced, the calculation time is shortened, and the high-precision and efficient simulation of the flow field is realized.     

Anisotropic black phosphorene nanotube anodes afford ultrafast kinetic rate or extra capacities for Li-ion batteries

As an important anode material for fast-charging Li-ion batteries (LIBs), black phosphorus (BP) has attracted extensive attention. Black phosphorene nanotubes (BPNTs) can be theoretically produced by rolling up the black phosphorene nanosheet along armchair (a-BPNTs) and zigzag (z-BPNTs) directions. The effects of curvature, chirality, Li-storage concentrations and strain stress on the Li-storage performance such as Li diffusion barriers and mechanical stabilities of BPNTs are mainly investigated by first principles calculations. The theoretical calculations predict that the a-BPNTs and z-BPNTs have good maximum Li-storage capacities, and the z-BPNTs exhibit better flexibility than a-BPNTs. The mechanical stabilities and Li-migration are all related to the curvature of BPNTs. Additionally, both a-BPNTs and z-BPNTs exhibit fast Li-ion conductivity along the c-axis direction. Moreover, the average Poisson's ratio of a-BPNTs (0.68) is larger than that of z-BPNTs (0.17), indicating that the strain stress is more difficult to apply on a-BPNTs than z-BPNTs. Our calculations predict that the a-BPNTs can afford ultrafast kinetic rate for fast-charging and high-power LIBs, while the z-BPNTs can provide extra capacity for high-energy LIBs.     

Metal-free regioselective construction of 2-aryl-2H-tetrazol-5-yl difluoromethylene phosphonates

Three bench-stable difluoromethylene phosphonate hydrazones were prepared from simple diethyl(difluoromethyl)phosphonate within two steps in good yields. The [3 + 2] cycloaddition reaction of these diazo precursors with aryl diazonium salts has been accomplished under metal-free conditions with exclusive regioselectivity. This transformation provides practical access to a broad panel of 2-aryl-2 H-tetrazol-5-yl difluoromethylene phosphonates, including the corresponding derivatives of amino acid...     

Semi-invariants of binary forms and Sylvester’s theorem

The Ramanujan Journal - We obtain a combinatorial formula related to the shear transformation for semi-invariants of binary forms, which implies the classical characterization of semi-invariants in...     

Vanishing coefficients and identities concerning Ramanujan’s parameters

The Ramanujan Journal - In this paper we investigate several infinite products with vanishing Taylor coefficients in arithmetic progressions. These infinite products are closely related to...