What is Behind the Meta-Learning Initialization of Adaptive Filter?-- a Naive Method for Accelerating Convergence of Adaptive Multichannel Active Noise Control

SHI, DONGYUAN; Gan, Woon-Seng; Shen, xiaoyi; Luo, Zhengding; Ji, Junwei

doi:10.2139/ssrn.4596210

Cited by 1 publication

(5 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Introduction of adaptive filtering layer: Unlike traditional graph convolutional networks that use shared weight matrices, the FEGNS framework designs adaptive filtering weights for each node, enabling a more sensitive handling of local features (Shi et al , 2024).…”

Section: Methodsmentioning

confidence: 99%

“…To accurately predict output locations while reducing computational demands, we introduce an adaptive filtering graph convolution network (AF-GCN). The AF-GCN processes input features from each node, X(i), and adaptively predicts output features, Y(i), as per equation (5) (Shi et al , 2024): The GNN process involves aggregating node vectors to amalgamate features from a node’s neighborhood, N ( i ), through a weighted sum with attentive weights, α ( i, j ), encapsulating local interactions. Dense layers elevate aggregated features into abstract representations, aligning with hierarchical feature learning.…”

Section: Methodsmentioning

confidence: 99%

“…These coefficients allow for precise weighting of different graph convolution branches. Through this design, the output feature of node i after adaptive filtering is computed as (Shi et al , 2024): …”

Section: Methodsmentioning

confidence: 99%

“…The equation F = Σfi represents summing individual forces fi acting on each particle, essential for accurately simulating overall fluid dynamics. Our approach sets FEGNS apart with two key innovations: adaptive filtering layer, an innovative technique that allows independent adjustment of filtering weights per node, enabling more accurate capture of local fluid dynamics features (Shi et al , 2024); and aggregator fusion strategy, which combines the results of various aggregation functions (e.g. mean, maximum) to significantly enhance the network’s expressive power in capturing complex fluid interactions. …”

Section: Introductionmentioning

confidence: 99%

“…adaptive filtering layer, an innovative technique that allows independent adjustment of filtering weights per node, enabling more accurate capture of local fluid dynamics features (Shi et al , 2024); and…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Efficient modeling of liquid splashing via graph neural networks with adaptive filter and aggregator fusion

Nan,

Feng,

et al. 2024

HFF

View full text Add to dashboard Cite

Purpose The purpose of this study is to advance the computational modeling of liquid splashing dynamics, while balancing simulation accuracy and computational efficiency, a duality often compromised in high-fidelity fluid dynamics simulations. Design/methodology/approach This study introduces the fluid efficient graph neural network simulator (FEGNS), an innovative framework that integrates an adaptive filtering layer and aggregator fusion strategy within a graph neural network architecture. FEGNS is designed to directly learn from extensive liquid splash data sets, capturing the intricate dynamics and intrinsically complex interactions. Findings FEGNS achieves a remarkable 30.3% improvement in simulation accuracy over traditional methods, coupled with a 51.6% enhancement in computational speed. It exhibits robust generalization capabilities across diverse materials, enabling realistic simulations of droplet effects. Comparative analyses and empirical validations demonstrate FEGNS’s superior performance against existing benchmark models. Originality/value The originality of FEGNS lies in its adaptive filtering layer, which independently adjusts filtering weights per node, and a novel aggregator fusion strategy that enriches the network’s expressive power by combining multiple aggregation functions. To facilitate further research and practical deployment, the FEGNS model has been made accessible on GitHub (https://github.com/nanjinyao/FEGNS/tree/main).

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%