This paper presents three different structures of index guided hexagonal shaped Hollow Core Photonic Crystal Fiber (HC-PCF) specially designed for sensing harmful food additives like Saccharin, Sorbitol and Butyl Acetate. The proposed PCFs have three distinct combinations of hexagonal and/or circular airholes at the innermost cladding layer and in the core region. A comparative study among the proposed structures reveals that the introduction of the hexagonal airholes at the innermost cladding layer increases the sensitivity. Also, the addition of the large circular airholes at the outermost layer reduces the confinement loss in all three structures. The diameters of the airholes of the innermost and outermost layer of the cladding region are gradually varied to optimize the PCFs. The numerical inquiry reveals that the optimized best structure shows relative sensitivity of 88.75%, 87.37% and 86.72% for Saccharin, Sorbitol, and Butyl Acetate respectively at the operating wavelength of 1.33 μm. The performances of the proposed structures are also investigated using Ethanol as the sensed sample for the purpose of comparison with previously reported works. The comparison shows that the introduced fibers outperform most of the recent works. Numerical analyses of the proposed structures are conducted using Full Vectorial Finite Element Method (FV-FEM).

Detection of harmful food additives using highly sensitive photonic crystal fiber

Si7N3 material filled novel heptagonal photonic crystal fiber for laser applications

Elliptical As2Se3 filled core ultra-high-nonlinearity and polarization-maintaining photonic crystal fiber with double hexagonal lattice cladding

A novel design of Hexa Sectored Photonic Crystal Fiber (HS-PCF) with high nonlinearity and high birefringence has been revealed in this paper where core is slotted and filled with Gallium Phosphide (GaP). Finite Element Method has been used for numerical investigation of the proposed PCF along with finer mesh. Different optical parameters like nonlinearity, effective area, power fraction, birefringence, confinement loss and Numerical Aperture (NA) have been explored by proper tuning of Geometrical variables. The investigation shows that, proposed PCF exhibits high nonlinearity of 9.47 × 104 W−1Km−1 at the operating wavelength of 1.4 µm along with high birefringence of 0.259, Numerical aperture of 0.8774 and very low confinement loss of 5.78 × 10−9 dB/m at the optical wavelength of 2.0 µm. Therefore, it is expected that this proposed PCF could be a strong candidate in biomedical imaging, super continuum generation and sensing applications considering polarized light.

Investigation of highly birefringent and highly nonlinear Hexa Sectored PCF with low confinement loss

A novel Decahedron photonic crystal fiber (DHn-PCF) is proposed. To sustain high birefringence and nonlinearity, the central elliptical core is filled with a highly nonlinear 2D material molybdenum disulphide (MoS2). The significant performance parameters of the proposed DHn-PCF like birefringence, nonlinearity, confinement loss, and core power fraction has been investigated with change in its structural parameters. Simulation results confirmed an exceptionally high birefringence and nonlinearity of 0.632 and 9.68 × 105 W−1 Km−1 with the outstanding core power fraction of 98.6% for the well-known communication window at 1.55 μ m . Besides, the effective material loss (EML) and confinement loss (CL) are also found very low as 0.0302 c m - 1 and 8.7 × 10 - 10 d B / m for the same communication window. Therefore, with all these excellent outcomes, the proposed DHn-PCF proves itself a potential candidate for polarization maintenance, residual polarization compensation, dispersion compensator, bio-medical imaging, supercontinuum and solitons generation, high bit data rate telecommunication and sensing application.

An ultra-high birefringent and nonlinear decahedron photonic crystal fiber employing molybdenum disulphide (MoS2): A numerical analysis

Dual-cladding high-birefringence and high-nonlinearity photonic crystal fiber with As2S3 core

Drug-Drug interactions (DDIs) are a challenging problem in drug research. Drug combination therapy is an effective solution to treat diseases, but it can also cause serious side effects. Therefore, DDIs prediction is critical in pharmacology. Recently, researchers have been using deep learning techniques to predict DDIs. However, these methods only consider single information of the drug and have shortcomings in robustness and scalability.In this paper, we propose a multi-type feature fusion based on graph neural network model (MFFGNN) for DDI prediction, which can effectively fuse the topological information in molecular graphs, the interaction information between drugs and the local chemical context in SMILES sequences. In MFFGNN, to fully learn the topological information of drugs, we propose a novel feature extraction module to capture the global features for the molecular graph and the local features for each atom of the molecular graph. In addition, in the multi-type feature fusion module, we use the gating mechanism in each graph convolution layer to solve the over-smoothing problem during information delivery. We perform extensive experiments on multiple real datasets. The results show that MFFGNN outperforms some state-of-the-art models for DDI prediction. Moreover, the cross-dataset experiment results further show that MFFGNN has good generalization performance.Our proposed model can efficiently integrate the information from SMILES sequences, molecular graphs and drug-drug interaction networks. We find that a multi-type feature fusion model can accurately predict DDIs. It may contribute to discovering novel DDIs. 

/pdf/multi-type-feature-fusion-based-on-graph-neural-network-for-18w0ufq3.pdf

Multi-type feature fusion based on graph neural network for drug-drug interaction prediction

/pdf/multi-type-feature-fusion-based-on-graph-neural-network-for-2b6xs167.pdf

There has been significant progress in skeleton-based action recognition. Human skeleton can be naturally structured into graph, so graph convolution networks have become the most popular method in this task. Most of these state-of-the-art methods optimized the structure of human skeleton graph to obtain better performance. Based on these advanced algorithms, a simple but strong network is proposed with three major contributions. Firstly, inspired by some adaptive graph convolution networks and non-local blocks, some kinds of self-attention modules are designed to exploit spatial and temporal dependencies and dynamically optimize the graph structure. Secondly, a light but efficient architecture of network is designed for skeleton-based action recognition. Moreover, a trick is proposed to enrich the skeleton data with bones connection information and make obvious improvement to the performance. The method achieves 90.5% accuracy on cross-subjects setting (NTU60), with 0.89M parameters and 0.32 GMACs of computation cost. This work is expected to inspire new ideas for the field. 

/pdf/an-efficient-self-attention-network-for-skeleton-based-3kah66bq.pdf

An efficient self-attention network for skeleton-based action recognition

/pdf/an-efficient-self-attention-network-for-skeleton-based-1srjsq3h.pdf

Xiaofei Qin

Papers

Dual-cladding high-birefringence and high-nonlinearity photonic crystal fiber with As2S3 core

Multi-type feature fusion based on graph neural network for drug-drug interaction prediction

Multi-type feature fusion based on graph neural network for drug-drug interaction prediction

An efficient self-attention network for skeleton-based action recognition

An efficient self-attention network for skeleton-based action recognition