A Distributed Neural Network Architecture for Robust Non-Linear Spatio-Temporal Prediction
This addresses the challenge of scalable and robust spatio-temporal prediction for applications like climate modeling or traffic forecasting, though it appears incremental as it builds on existing neural network concepts.
The paper tackles the problem of predicting complex spatio-temporal dynamics by introducing DISTANA, a distributed neural network architecture that learns in parallel and scales to large spaces, achieving robustness to overfitting compared to other ANN models.
We introduce a distributed spatio-temporal artificial neural network architecture (DISTANA). It encodes mesh nodes using recurrent, neural prediction kernels (PKs), while neural transition kernels (TKs) transfer information between neighboring PKs, together modeling and predicting spatio-temporal time series dynamics. As a consequence, DISTANA assumes that generally applicable causes, which may be locally modified, generate the observed data. DISTANA learns in a parallel, spatially distributed manner, scales to large problem spaces, is capable of approximating complex dynamics, and is particularly robust to overfitting when compared to other competitive ANN models. Moreover, it is applicable to heterogeneously structured meshes.