scholarly journals A new multilayer graph model for speech signals with graph learning

2021 ◽  
pp. 103360
Author(s):  
Tingting Wang ◽  
Haiyan Guo ◽  
Qiquan Zhang ◽  
Zhen Yang
2021 ◽  
Vol 22 (1) ◽  
Author(s):  
Jinlong Hu ◽  
Lijie Cao ◽  
Tenghui Li ◽  
Shoubin Dong ◽  
Ping Li

Abstract Background Autism spectrum disorders (ASD) imply a spectrum of symptoms rather than a single phenotype. ASD could affect brain connectivity at different degree based on the severity of the symptom. Given their excellent learning capability, graph neural networks (GNN) methods have recently been used to uncover functional connectivity patterns and biological mechanisms in neuropsychiatric disorders, such as ASD. However, there remain challenges to develop an accurate GNN learning model and understand how specific decisions of these graph models are made in brain network analysis. Results In this paper, we propose a graph attention network based learning and interpreting method, namely GAT-LI, which learns to classify functional brain networks of ASD individuals versus healthy controls (HC), and interprets the learned graph model with feature importance. Specifically, GAT-LI includes a graph learning stage and an interpreting stage. First, in the graph learning stage, a new graph attention network model, namely GAT2, uses graph attention layers to learn the node representation, and a novel attention pooling layer to obtain the graph representation for functional brain network classification. We experimentally compared GAT2 model’s performance on the ABIDE I database from 1035 subjects against the classification performances of other well-known models, and the results showed that the GAT2 model achieved the best classification performance. We experimentally compared the influence of different construction methods of brain networks in GAT2 model. We also used a larger synthetic graph dataset with 4000 samples to validate the utility and power of GAT2 model. Second, in the interpreting stage, we used GNNExplainer to interpret learned GAT2 model with feature importance. We experimentally compared GNNExplainer with two well-known interpretation methods including Saliency Map and DeepLIFT to interpret the learned model, and the results showed GNNExplainer achieved the best interpretation performance. We further used the interpretation method to identify the features that contributed most in classifying ASD versus HC. Conclusion We propose a two-stage learning and interpreting method GAT-LI to classify functional brain networks and interpret the feature importance in the graph model. The method should also be useful in the classification and interpretation tasks for graph data from other biomedical scenarios.


2008 ◽  
Vol 1 (06) ◽  
pp. 329-334
Author(s):  
S. Rabih ◽  
C. Turpin ◽  
S. Astier

Author(s):  
A. N. Bozhko

Computer-aided design of assembly processes (Computer aided assembly planning, CAAP) of complex products is an important and urgent problem of state-of-the-art information technologies. Intensive research on CAAP has been underway since the 1980s. Meanwhile, specialized design systems were created to provide synthesis of assembly plans and product decompositions into assembly units. Such systems as ASPE, RAPID, XAP / 1, FLAPS, Archimedes, PRELEIDES, HAP, etc. can be given, as an example. These experimental developments did not get widespread use in industry, since they are based on the models of products with limited adequacy and require an expert’s active involvement in preparing initial information. The design tools for the state-of-the-art full-featured CAD/CAM systems (Siemens NX, Dassault CATIA and PTC Creo Elements / Pro), which are designed to provide CAAP, mainly take into account the geometric constraints that the design imposes on design solutions. These systems often synthesize technologically incorrect assembly sequences in which known technological heuristics are violated, for example orderliness in accuracy, consistency with the system of dimension chains, etc.An AssemBL software application package has been developed for a structured analysis of products and a synthesis of assembly plans and decompositions. The AssemBL uses a hyper-graph model of a product that correctly describes coherent and sequential assembly operations and processes. In terms of the hyper-graph model, an assembly operation is described as shrinkage of edge, an assembly plan is a sequence of shrinkages that converts a hyper-graph into the point, and a decomposition of product into assembly units is a hyper-graph partition into sub-graphs.The AssemBL solves the problem of minimizing the number of direct checks for geometric solvability when assembling complex products. This task is posed as a plus-sum two-person game of bicoloured brushing of an ordered set. In the paradigm of this model, the brushing operation is to check a certain structured fragment for solvability by collision detection methods. A rational brushing strategy minimizes the number of such checks.The package is integrated into the Siemens NX 10.0 computer-aided design system. This solution allowed us to combine specialized AssemBL tools with a developed toolkit of one of the most powerful and popular integrated CAD/CAM /CAE systems.


Author(s):  
Yoshito Mekada ◽  
Miyuki Mukasa ◽  
Hiroshi Hasegawa ◽  
Masao Kasuga ◽  
Shuichi Matsumoto ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document