Search

Highlights • An intermediate model is proposed to support global performance evaluation. • Cross-parameterization algorithm is adopted to compute intermediate results. • Two feature selection methods are adopted to improve prediction accuracy. • Machine learning based approach is applied to realize global prediction. • Extensive experiments are conducted for performance verification. Abstract Historical simulation data reuse is crucial for helping the designer improve the product development process. Currently, simulation data mining has been brought into use to discover the underlying knowledge to support efficient design changes. However, most of the existing simulation data mining methods paid little attention to global performance evaluation, and thus causing it difficult for the designer to browse all the simulation results conveniently and accurately if it is without actual simulation performance verification. In this study, a machine learning based global simulation data mining approach is proposed to discover the interrelations between key design parameters and global performance parameters to realize the accurate prediction of all the simulation results, and thus supporting the decision-making in the development process. Firstly, an intermediate mesh model based cross-parameterization algorithm is adopted to construct global performance evaluation indicators. After that, two feature selection methods for design parameters are applied to select salient single parameter and their combinations to reduce the modeling complexity and improve the prediction accuracy. Finally, a machine learning based simulation data mining approach is developed and improved to realize global performance evaluation accurately and efficiently. Extensive experiments are conducted to demonstrate the feasibility, effectiveness and correctness of the proposed approach. [ABSTRACT FROM AUTHOR]

Nephrotoxicity is a main problem in cancer patients using cisplatin. Oxidative stress, inflammation and apoptosis are the important mechanisms of cisplatin induced nephrotoxicity. In the present study, we investigated the effect of the extracts of morning glory on nephrotoxicity by cisplatin in human embryonic kidney cells 293 (HEK-293) and mice. Previous studies have reported that morning glory extracts showed potent activity on anti-inflammatory and anti-oxidant. However, the protective effects of the n -hexane layer of morning glory seed (MGs-Hx) on nephrotoxicity and its mechanisms have not been clearly understood. Oral administration with MGs-Hx showed protective effects in vivo experiments test and the treatment of MGs-Hx in a concentration of 100 mg/kg/day had significant effect both of decreasing serum creatinine, BUN, serum uric acid level and reduced iNOS, COX-2 mRNA expressions with low side-effect. Moreover, cell viability was restored by MGs-Hx treatment compared to cisplatin-induced nephrotoxic HEK-293 cells. Co-treatment with MGs-Hx and cisplatin showed the significant effect to reduce inflammatory enzyme, iNOS expression and continuous production of NO. In addition, it exhibited a tendency to decreasing expression of apoptosis-related proteins, caspase-3, 8 and 9, and NF-κB translocation to nucleus as well as phosphorylation of p38, JNK, ERK in cisplatin-induced nephrotoxic HEK-293 cells. Our study provides insight into the underlying mechanisms of MGs-Hx and suggests that MGs-Hx might be a potential therapeutic agent to modulate inflammation and apoptosis in nephrotoxicity. [ABSTRACT FROM AUTHOR]

Model simplification is imperative in the process of computer aided design (CAD) and computer aided engineering (CAE) integration. Mid-surface abstraction is the most effective method to simplify the thin-walled models. Many previous research efforts have been focused on the mid-surface abstraction, including the model decomposition based methods, Medial Axis Transform (MAT) based methods and Chordal Axis Transform (CAT) based methods. However, complex thin-walled models cannot be handled well due to the fact that there are some problems including low geometrical precision, poor topological structure, etc., in the above resultant mid-surface models. Especially, these methods are hard to be reused to generate the mid-surface model efficiently. Therefore, a hierarchical semantic mid-surface abstraction method is proposed for the thin-walled model based on rib feature decomposition in this paper. Firstly, a new hierarchical semantic structure is defined and applied on both the thin-walled models and mid-surface models. After that, the model decomposition is conducted based on the identified rib features and the hierarchical semantic information is obtained at the same time. Then, the offset operation and discretization based methods are used to obtain the mid-surface patch for each sub region with different semantic structure respectively. Finally, the mid-surface model with hierarchical semantic information is generated by stitching all discrete patches. Moreover, the above model can be reused to facilitate the rapid mid-surface abstraction of the changed model based on its hierarchical semantic structure. Several examples are given to demonstrate the outperformance of the proposed method. [ABSTRACT FROM AUTHOR]

The efficient and integrated visualization and reuse of multidisciplinary simulation data are imperative for the development of complex products. However, it is not a trivial task for designers to efficiently acquire, view and then reuse simulation data to improve the product design process. The challenge is that simulation data are always too huge to be transferred and retrieved quickly and they tend to be heterogeneous and tool-specific due to the lack of a uniform representation. In this study, an approach of intermediate model based efficient and integrated visualization of multidisciplinary simulation data for simulation information reuse is proposed to address the above issues. Firstly, the intermediate model based integrated model framework is designed to support the uniform modeling and integrated visualization of multidisciplinary simulation data. Then, the intermediate mesh model is constructed based on the hybrid mesh size (HMS) field to achieve the uniform representation of multidisciplinary simulation data with high-fidelity. Thirdly, a series of strategies such as coarse filtering, fine filtering for incremental transmission and data compression are proposed to improve data transmission efficiency. Moreover, a simulation model descriptor (SMD) based similarity assessment method is developed to support the efficient retrieval of simulation models for reuse. Finally, several experiments are conducted to demonstrate the feasibility and effectiveness of the proposed model and methods. [ABSTRACT FROM AUTHOR]

The recurrent network architecture is a widely used model in sequence modeling, but its serial dependency hinders the computation parallelization, which makes the operation inefficient. The same problem was encountered in serial adder at the early stage of digital electronics. In this paper, we discuss the similarities between recurrent neural network (RNN) and serial adder. Inspired by carry-lookahead adder, we introduce carry-lookahead module to RNN, which makes it possible for RNN to run in parallel. Then, we design the method of parallel RNN computation, and finally Carry-lookahead RNN (CL-RNN) is proposed. CL-RNN takes advantages in parallelism and flexible receptive field. Through a comprehensive set of tests, we verify that CL-RNN can perform better than existing typical RNNs in sequence modeling tasks which are specially designed for RNNs. Code and models are available at: https://github.com/WinnieJiangHW/Carry-lookahead_RNN. [ABSTRACT FROM AUTHOR]

Performing 3-dimensional computer-aided design (3D-CAD) model classification, retrieval, and reuse is of vital importance in industrial manufacturing, as it considerably shortens the engineering development cycle and reduces development costs. Although existing 3D model classification and retrieval methods achieve satisfactory performance when operating on meshes or point clouds, they cannot be applied directly to 3D-CAD models, which are generally represented by the boundary representation (B-rep). To address this issue, and to fully exploit the topology of B-rep, a graph structure descriptor called B-rep graph is proposed to pre-process B-rep data, and to extract the precise topological and geometric features from 3D-CAD models. Meanwhile, a novel efficient neural network called FuS-GCN, based on graph convolutional networks (GCNs), is designed to handle this graph data. To better extract the graph features and to improve the effect of pooling, the self-attention mechanism and feature fusion are incorporated into the pooling layer, yielding the proposed fusion self-attention graph pooling (FuSPool) algorithm. Finally, we demonstrate the effectiveness of FuS-GCN on 3D-CAD model data, while outperforming alternative 3D shape descriptors such as point clouds, voxels, and meshes. [ABSTRACT FROM AUTHOR]

Current geometric optimization methods do not fully consider the overall structure of the model, which may make the effect of geometric optimization of the hex meshes for the models with concave curves unsatisfactory. To this end, this paper proposes an approach to structure-aware geometric optimization of hex meshes. In the approach, by relocating the position of the base complex of the hex mesh, the overall parameterized energy is reduced to obtain an optimized geometric embedding. First, a frame field which conforms to the base complex structure of the input hex mesh is generated. Then, a parameterization that satisfies the structure constraints is established based on the frame field. Finally, the positions of the singular lines of the hex mesh are optimized according to the gradient of the parameterization energy function, and the key isoparametric surfaces are used as the position constraints of the vertices of the patches of the base complex to optimize the whole hex mesh. The experimental results show that the scaled Jacobian value of hex meshes can be effectively improved by the proposed method. • The parameterization energy is reduced to obtain an optimized geometric embedding. • The singular lines are relocated to achieve a less global volume distortion. • The patches are optimized on the extracted isoparametric surfaces. • A topologically compatible frame field is designed to support the parameterization. [ABSTRACT FROM AUTHOR]