Ыстық алгоритмдер
method
ID: greedyalgorithms
Related Papers
Dependency Parsing with Backtracking using Deep Reinforcement Learning
AbstractGreedy algorithms for NLP such as transition-based parsing are prone to error propagation. One way to overcome this problem is to allow the algorithm to backtrack and explore an alternative …...