Interpreting Temporal Graph Neural Networks with Koopman Theory
UiT The Arctic University of Norway, Department of Mathematics and Statistics(挪威北极大学数学与统计系);Sapienza Università di Roma, Department of Information Engineering, Electronics and Telecommunications(罗马大学信息工程、电子与电信系);NORCE Norwegian Research Centre AS(挪威研究中心AS)
Michele Guerra, Simone Scardapane, Filippo Maria Bianchi
Spatiotemporal graph neural networks (STGNNs) have shown promising results in many domains, from forecasting to epidemiology. However, understanding the dynamics learned by these models and explaining their behaviour is significantly more difficult than for models that deal with static data. Inspired by Koopman theory, which allows a simple description of intricate, nonlinear dynamical systems, we introduce new explainability approaches for temporal graphs. Specifically, we present two methods to interpret the STGNN's decision process and identify the most relevant spatial and temporal patterns in the input for the task at hand. The first relies on dynamic mode decomposition (DMD), a Koopman-inspired dimensionality reduction method. The second relies on sparse identification of nonlinear dynamics (SINDy), a popular method for discovering governing equations of dynamical systems, which we use for the first time as a general tool for explainability. On semi-synthetic dissemination datasets, our methods correctly identify interpretable features such as the times at which infections occur and the infected nodes. We also validate the methods qualitatively on a real-world human motion dataset, where the explanations highlight the body parts most relevant for action recognition.
School of Electrical and Mechanical Engineering, The University of Adelaide(电气与机械工程学院,阿德莱德大学);Information Systems Technology and Design Pillar, Singapore University of Technology and Design(信息系统技术与设计支柱,新加坡科技设计大学)
Temporally Consistent Koopman Autoencoders for Forecasting Dynamical Systems
ElectroScience Laboratory, The Ohio State University(电科学实验室,俄亥俄州立大学);Department of Electrical and Computer Engineering, The Ohio State University(电气与计算机工程系,俄亥俄州立大学);SLAC National Accelerator Laboratory, Stanford University(斯坦福大学SLAC国家加速器实验室)
Transformer with Koopman-Enhanced Graph Convolutional Network for Spatiotemporal Dynamics Forecasting
Charles F. Dolan School of Business, Fairfield University(Fairfield大学商学院);Department of Industrial & Systems Engineering, The University of Tennessee(田纳西大学工业与系统工程系)
Toward Scalable SDN for LEO Mega-Constellations: A Graph Learning Approach
School of Electrical and Mechanical Engineering, University of Adelaide(阿德莱德大学电气与机械工程学院);College of Computing and Informatics, University of Sharjah(沙迦大学计算机与信息学院);Singapore University of Technology and Design(新加坡科技与设计大学);Electronics and Telecommunications Research Institute(电子电信研究所)
本文提出一种可扩展的分层 SDN 框架,利用图神经网络和 Koopman 理论来优化低轨 mega-星座的网络管理,实验证明在空间压缩和时间预测方面有显著提升。
Recurrent neural networks and Koopman-based frameworks for temporal predictions in a low-order model of turbulence
University of Tehran(德黑兰大学);KTH Royal Institute of Technology(瑞典皇家理工学院);Swedish e-Science Research Centre (SeRC)(瑞典e-Science研究中心)
Moehlis等人(New J. Phys. 6,56,2004)在预测近壁湍流低阶模型的时间动力学中评估了循环神经网络和基于Koopman的框架的能力。我们的研究结果表明,有可能获得良好的长期统计和混沌系统的动态行为的再现与适当训练的长短期记忆(LSTM)网络,导致相对误差的平均值和波动低于$1\%$。此外,一个新开发的基于Koopman的框架,称为Koopman与非线性强迫(KNF),导致相同水平的准确性的统计数据在一个显着降低计算费用。此外,KNF框架在短期预测方面优于LSTM网络。我们还观察到,使用的损失函数的基础上的混沌系统的瞬时预测可以导致次优再现长期统计。因此,我们提出了一个模型选择标准的基础上计算的统计,它允许实现良好的统计重建,即使在小的数据集,在瞬时预测的准确性损失最小。
P-K-GCN: Physics-augmented Koopman-enhanced Graph Convolutional Network for Deep Spatiotemporal Super-resolution
Department of Industrial & Systems Engineering, The University of Tennessee, Knoxville, TN, USA(工业与系统工程系,田纳西大学, Knoxville,TN,美国);Charles F. Dolan School of Business, Fairfield University, Fairfield, USA(查尔斯·F·多兰商学院,费尔菲尔德大学, Fairfield,美国);Department of Electrical Engineering & Computer Science, The University of Tennessee, Knoxville, TN, USA(电气工程与计算机科学系,田纳西大学, Knoxville,TN,美国)
DeepKoopFormer: A Koopman Enhanced Transformer Based Architecture for Time Series Forecasting
Helmholtz Center for Environmental Research-UFZ(海德堡环境研究中心-UFZ);Delft Center for Systems and Control(代尔夫特系统与控制中心);School of Engineering, University of Pittsburgh(匹兹堡大学工程学院)
Semantic Communication and Control Co-Design for Multi-Objective Distinct Dynamics
Centre for Wireless Communications, University of Oulu(无线通信中心,奥卢大学);Department of Electronics and Communications Engineering, Kwangwoon University(电子与通信工程系,光州大学)
Learning Generalizable Reconstruction of High-Dimensional Neural Dynamics
Heidelberg University(海德堡大学);Interdisciplinary Centre for Scientific Computing (IWR)(跨学科科学计算中心(IWR));Department of Mathematics and Computer Science(数学与计算机科学系)
FreKoo++: Learning Continuous Spectral Dynamics for Temporal Domain Generalization
Australian Artificial Intelligence Institute (AAII)(澳大利亚人工智能研究所(AAII));Faculty of Engineering and Information Technology(工程与信息技术学院);University of Technology Sydney (UTS)(悉尼科技大学(UTS))
Certified Neural Approximations of Nonlinear Dynamics
Delft Center for Systems and Control(代尔夫特系统与控制中心);Department of Computer Science, University of Oxford(牛津大学计算机科学系);The Italian Institute of Artificial Intelligence (AI4I)(意大利人工智能研究所(AI4I))
EEG-SeeGraph: Interpreting functional connectivity disruptions in dementias via sparse-explanatory dynamic EEG-graph learning
Harbin Institute of Technology, Shenzhen, China(哈尔滨工业大学(深圳));Institute of Neurological Diseases, Shenzhen Bay Laboratory, Shenzhen, China(深圳湾实验室神经疾病研究院)