id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
cord-188465-wwi8uydi	Spadon, Gabriel	Pay Attention to Evolution: Time Series Forecasting with Deep Graph-Evolution Learning	2020-08-28		.txt	text/plain	10009	517	57	Definition ω ∈ N + Sliding window size w, z ∈ N + Number of training and testing (i.e., stride) timestamps s, t, v ∈ N + Number of samples, timestamps, and variables T ∈ R s×t×v Tensor of multiple multivariate time-series Y ∈ R s×ω×v Batched input of the first GSE and the Autoregression layers Yα ∈ R s×ω×v Output of the first GSE and input of the encoder layers Yε ∈ R s×ω×v Output of the encoder and input of the decoder layers Yε ∈ R s×z×v Output from the first recurrent unit and input to the second one Y ∈ R s×z×v Output of the second recurrent unit and input of the second GSE layer Y ψ ∈ R s×z×v Non-linear output yielded by the second GSE layer Y λ ∈ R s×z×v Linear output provided by the Autoregression layer Y ∈ R s×z×v Final result from the merging of the linear and non-linear outputs G = V, E Graph in which V is the set of nodes and E the set of edges A ∈ R v×v Adjacency matrix of co-occurring variables Aµ ∈ R v×v Adjacency matrix shared between GSE layers A φ ∈ R v×v Evolved adjacency matrix produced by the second GSE layer U • V Batch-wise Hadamard product between matrices U and V U · V Batch-wise scalar product between matrices U and V · F	./cache/cord-188465-wwi8uydi.txt	./txt/cord-188465-wwi8uydi.txt