Attentive Neural Controlled Differential Equations for Time-series Classification and Forecasting

Jhin, Sheo Yon; Heejoo, Shin,; Hong, Seoyoung; Jo, Minju; Park, Solhee; Park, Noseong; Lee, Seungbeom; Maeng, Hwiyoung; Jeon, Seungmin

doi:10.1109/icdm51629.2021.00035

Cited by 9 publications

(4 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to establish a relevant connection between K and W S1 , we adopt a node embedding matrix E 1 that controls both positive and negtive relations and get weight pool of each node with W S1 . To enhance the sparsity of the graph, we employ an adaptive mask matrix (Jhin et al 2021b). Formally,…”

Section: Methodsmentioning

confidence: 99%

“…In recent years, a noteworthy trend has emerged, involving the fusion of ordinary differential equations (ODEs) with neural networks to acquire insights into continuous-time dynamics (Chen et al 2018;Rubanova, Chen, and Duvenaud 2019;Kidger et al 2020;Jhin et al 2021a;Hwang et al 2021;Huang, Sun, and Wang 2021;Fang et al 2021;Choi et al 2022;Jhin et al 2022). These hybrids, encompassing both ODEs and graph neural networks (GNNs), have demonstrated promising performance across a variety of tasks.…”

Section: Introductionmentioning

confidence: 99%

“…These hybrids, encompassing both ODEs and graph neural networks (GNNs), have demonstrated promising performance across a variety of tasks. This includes climate modeling (Hwang et al 2021;Jhin et al 2021a), traffic flow prediction (Poli et al 2019;Fang et al 2021;Choi et al 2022), node classification (Zang and Wang 2020;Xhonneux, Qu, and Tang 2020), and dynamic interactions (Huang, Sun, and Wang 2021). However, it remains pertinent to note that a substantial portion of complex networks retains enigmatic dynamics yet to be fully unraveled.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Signed Graph Neural Ordinary Differential Equation for Modeling Continuous-Time Dynamics

Chen,

Wu,

Lou

et al. 2024

AAAI

View full text Add to dashboard Cite

Modeling continuous-time dynamics constitutes a foundational challenge, and uncovering inter-component correlations within complex systems holds promise for enhancing the efficacy of dynamic modeling. The prevailing approach of integrating graph neural networks with ordinary differential equations has demonstrated promising performance. However, they disregard the crucial signed information potential on graphs, impeding their capacity to accurately capture real-world phenomena and leading to subpar outcomes. In response, we introduce a novel approach: a signed graph neural ordinary differential equation, adeptly addressing the limitations of miscapturing signed information. Our proposed solution boasts both flexibility and efficiency. To substantiate its effectiveness, we seamlessly integrate our devised strategies into three preeminent graph-based dynamic modeling frameworks: graph neural ordinary differential equations, graph neural controlled differential equations, and graph recurrent neural networks. Rigorous assessments encompass three intricate dynamic scenarios from physics and biology, as well as scrutiny across four authentic real-world traffic datasets. Remarkably outperforming the trio of baselines, empirical results underscore the substantial performance enhancements facilitated by our proposed approach. Our code can be found at https://github.com/beautyonce/SGODE.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Signed Graph Neural Ordinary Differential Equation for Modeling Continuous-Time Dynamics

Chen,

Wu,

Lou

et al. 2024

AAAI

View full text Add to dashboard Cite

show abstract

“…In (Jhin et al 2021b), they proposed ANCDEs to insert an attention mechanism into NCDEs. Unlike our method, however, their goal is to pick useful information from the path X.…”

Section: Related Work and Preliminariesmentioning

confidence: 99%

Learnable Path in Neural Controlled Differential Equations

Jhin

Kook

et al. 2023

AAAI

View full text Add to dashboard Cite

Neural controlled differential equations (NCDEs), which are continuous analogues to recurrent neural networks (RNNs), are a specialized model in (irregular) time-series processing. In comparison with similar models, e.g., neural ordinary differential equations (NODEs), the key distinctive characteristics of NCDEs are i) the adoption of the continuous path created by an interpolation algorithm from each raw discrete time-series sample and ii) the adoption of the Riemann--Stieltjes integral. It is the continuous path which makes NCDEs be analogues to continuous RNNs. However, NCDEs use existing interpolation algorithms to create the path, which is unclear whether they can create an optimal path. To this end, we present a method to generate another latent path (rather than relying on existing interpolation algorithms), which is identical to learning an appropriate interpolation method. We design an encoder-decoder module based on NCDEs and NODEs, and a special training method for it. Our method shows the best performance in both time-series classification and forecasting.

show abstract

Satellite State Prediction and Maneuver Detection Analysis Using NCDEs

Lee,

Woo

2024

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Attentive Neural Controlled Differential Equations for Time-series Classification and Forecasting

Cited by 9 publications

References 21 publications

Signed Graph Neural Ordinary Differential Equation for Modeling Continuous-Time Dynamics

Signed Graph Neural Ordinary Differential Equation for Modeling Continuous-Time Dynamics

Learnable Path in Neural Controlled Differential Equations

Satellite State Prediction and Maneuver Detection Analysis Using NCDEs

Contact Info

Product

Resources

About