Automatic Unit Test Data Generation and Actor-Critic Reinforcement Learning for Code Synthesis.
Philip John GorinskiMatthieu ZimmerGerasimos LampourasDerrick-Goh-Xin DeikIgnacio IacobacciPublished in: EMNLP (Findings) (2023)
Keyphrases
- actor critic
- reinforcement learning
- test data generation
- temporal difference
- policy gradient
- reinforcement learning algorithms
- function approximation
- approximate dynamic programming
- neuro fuzzy
- optimal control
- test cases
- gradient method
- policy iteration
- simulated annealing algorithm
- state space
- learning algorithm
- evaluation function
- model free
- policy gradient methods
- neural network
- markov decision processes
- optimal policy
- source code
- action selection
- software testing
- linear program
- temporal difference learning
- object oriented
- dynamic programming
- multi agent systems
- genetic algorithm