OptiGAN: Generative Adversarial Networks for Goal Optimized Sequence Generation
This addresses the challenge of goal-optimized sequence generation for tasks like text and real-valued sequences, though it appears incremental as it builds on existing GAN and RL methods.
The paper tackled the problem of generating sequences with specific desired goals, which current models do not optimize directly, by introducing OptiGAN, a model combining GANs and RL to optimize goal scores using policy gradients, achieving higher desired scores than baselines without sacrificing diversity.
One of the challenging problems in sequence generation tasks is the optimized generation of sequences with specific desired goals. Current sequential generative models mainly generate sequences to closely mimic the training data, without direct optimization of desired goals or properties specific to the task. We introduce OptiGAN, a generative model that incorporates both Generative Adversarial Networks (GAN) and Reinforcement Learning (RL) to optimize desired goal scores using policy gradients. We apply our model to text and real-valued sequence generation, where our model is able to achieve higher desired scores out-performing GAN and RL baselines, while not sacrificing output sample diversity.