"Neuromathematics" - Open platform to accelerate complex applied math problems solutions using neural network algorithms.
Pavel V. SkribtsovPavel A. KazantsevSvetlana V. KorobkovaDmitry V. PantiukhinIlia M. VoronkovAlexander I. GalushkinGeorgiy P. OstapenkoArtem M. LodyaginPublished in: IJCNN (2007)
Keyphrases
- neural network
- benchmark problems
- approximate solutions
- optimization problems
- specific problems
- search methods
- practical solutions
- partial solutions
- computationally intensive
- neural network is trained
- greedy algorithms
- real world
- difficult problems
- mathematical tools
- efficient solutions
- similar problems
- heuristic methods
- solution quality
- problems in computer vision
- exact and approximate
- computational issues
- solving complex
- optimization criteria
- exact algorithms
- run times
- stochastic local search algorithms
- data structure
- provide efficient solutions
- practical problems
- fuzzy logic
- exact solution
- test problems
- artificial neural networks
- combinatorial problems
- computationally expensive
- combinatorial optimization
- computational cost
- worst case
- simulated annealing
- significant improvement
- data mining techniques
- np complete
- genetic algorithm
- optimal or near optimal
- computationally efficient
- learning algorithm
- back propagation
- decision trees
- computational geometry
- convergence rate
- computational complexity
- search algorithm
- real time
- test instances
- related problems