Abstract
We introduce the notion of Hypergraph Weighted Model (HWM) that generically associates a tensor network to a hypergraph and then computes a value by tensor contractions directed by its hyperedges. A series r defined on a hypergraph family is said to be recognizable if there exists a HWM that computes it. This model generalizes the notion of recognizable series on strings and trees. We present some properties of the model and study at which conditions finite support series are recognizable.
Cite
CITATION STYLE
Bailly, R., Denis, F., & Rabusseau, G. (2015). Recognizable series on hypergraphs. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8977, pp. 639–651). Springer Verlag. https://doi.org/10.1007/978-3-319-15579-1_50
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
