Cracking nuts with a sledgehammer: when modern graph neural networks do worse than classical greedy algorithms.
Maria Chiara AngeliniFederico Ricci-TersenghiPublished in: CoRR (2022)
Keyphrases
- greedy algorithms
- neural network
- greedy algorithm
- pattern recognition
- genetic algorithm
- graph theory
- knapsack problem
- graph representation
- graph model
- neural network model
- graph structure
- fuzzy logic
- connected components
- directed graph
- search algorithm
- reinforced concrete
- graph databases
- multi layer
- weighted graph
- bipartite graph
- neural nets
- random walk
- worst case
- artificial neural networks
- dynamic programming
- radial basis function
- back propagation
- graph partitioning
- graph theoretic
- bayesian networks
- dependency graph