Performance of Graph Neural Networks for Point Cloud Applications

被引:0
|
作者
Parikh, Dhruv [1 ]
Zhang, Bingyi [1 ]
Kannan, Rajgopal [2 ]
Prasanna, Viktor [1 ]
Busart, Carl [2 ]
机构
[1] Univ Southern Calif, Los Angeles, CA 90007 USA
[2] DEVCOM US Army Res Lab, Adelphi, MD USA
关键词
Graph neural network; point cloud; k-nearest neighbors; dynamic graph construction; performance profiling;
D O I
10.1109/HPEC58863.2023.10363595
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Graph Neural Networks (GNNs) have gained significant momentum recently due to their capability to learn on unstructured graph data. Dynamic GNNs (DGNNs) are the current state-of-the-art for point cloud applications; such applications (viz. autonomous driving) require real-time processing at the edge with tight latency and memory constraints. Conducting performance analysis on such DGNNs, thus, becomes a crucial task to evaluate network suitability. This paper presents a profiling analysis of EdgeConv-based DGNNs applied to point cloud inputs. We assess their inference performance in terms of end-to-end latency and memory consumption on state-of-the-art CPU and GPU platforms. The EdgeConv layer has two stages: (1) dynamic graph generation using k-Nearest Neighbors (kNN) and, (2) node feature updation. The addition of dynamic graph generation via kNN in each (EdgeConv) layer enhances network performance compared to networks that work with the same static graph in each layer; such performance enhancement comes, however, at the added computational cost associated with the dynamic graph generation stage (via kNN algorithm). Understanding its costs is essential for identifying the performance bottleneck and exploring potential avenues for hardware acceleration. To this end, this paper aims to shed light on the performance characteristics of EdgeConv-based DGNNs for point cloud inputs. Our performance analysis on a state-of-the-art EdgeConv network for classification shows that the dynamic graph construction via kNN takes up upwards of 95% of network latency on the GPU and almost 90% on the CPU. Moreover, we propose a quasi-Dynamic Graph Neural Network (qDGNN) that halts dynamic graph updates after a specific depth within the network to significantly reduce the latency on both CPU and GPU whilst matching the original networks inference accuracy.
引用
收藏
页数:7
相关论文
共 50 条
  • [31] Hyperbolic Graph Neural Networks: A Tutorial on Methods and Applications
    Zhou, Min
    Yang, Menglin
    Xiong, Bo
    Xiong, Hui
    King, Irwin
    PROCEEDINGS OF THE 29TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, KDD 2023, 2023, : 5843 - 5844
  • [32] A Review of Graph Neural Networks and Their Applications in Power Systems
    Liao, Wenlong
    Bak-Jensen, Birgitte
    Pillai, Jayakrishnan Radhakrishna
    Wang, Yuelong
    Wang, Yusen
    JOURNAL OF MODERN POWER SYSTEMS AND CLEAN ENERGY, 2022, 10 (02) : 345 - 360
  • [33] A Review of Graph Neural Networks and Their Applications in Power Systems
    Wenlong Liao
    Birgitte Bak-Jensen
    Jayakrishnan Radhakrishna Pillai
    Yuelong Wang
    Yusen Wang
    Journal of Modern Power Systems and Clean Energy, 2022, 10 (02) : 345 - 360
  • [34] Edge Enhancement Oriented Graph Convolutional Networks for Point Cloud Segmentation
    Zhang, Xiaoyan
    Feng, Lin
    IEEE ACCESS, 2024, 12 : 70550 - 70558
  • [35] Leveraging graph neural networks for point-of-interest recommendations
    Zhang, Jiyong
    Liu, Xin
    Zhou, Xiaofei
    Chu, Xiaowen
    NEUROCOMPUTING, 2021, 462 : 1 - 13
  • [36] Explainability-Aware One Point Attack for Point Cloud Neural Networks
    Tan, Hanxiao
    Kotthaus, Helena
    2023 IEEE/CVF WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV), 2023, : 4570 - 4579
  • [37] GAPointNet: Graph attention based point neural network for exploiting local feature of point cloud
    Chen, Can
    Fragonara, Luca Zanotti
    Tsourdos, Antonios
    NEUROCOMPUTING, 2021, 438 : 122 - 132
  • [38] Encoding Partial Point Cloud Neighborhoods for Convolutional Neural Networks
    Chakraborty T.
    Krishnamurthy H.
    Computer-Aided Design and Applications, 2023, 20 (02): : 290 - 305
  • [39] Knowledge Enhanced Neural Networks for Point Cloud Semantic Segmentation
    Grilli, Eleonora
    Daniele, Alessandro
    Bassier, Maarten
    Remondino, Fabio
    Serafini, Luciano
    REMOTE SENSING, 2023, 15 (10)
  • [40] Spherical Fractal Convolutional Neural Networks for Point Cloud Recognition
    Rao, Yongming
    Lu, Jiwen
    Zhou, Jie
    2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019), 2019, : 452 - 460