Partial Off-policy Learning: Balance Accuracy and Diversity for Human-Oriented Image Captioning

Human-oriented image captioning with both high diversity and accuracy is a challenging task in vision+language modeling. The reinforcement learning (RL) based frameworks promote the accuracy of image captioning, yet seriously hurt the diversity. In contrast, other methods based on variational auto-encoder (VAE) or generative adversarial network (GAN) can produce diverse yet less accurate captions. In this work, we devote our attention to promote the diversity of RL-based image captioning. To be specific, we devise a partial off-policy learning scheme to balance accuracy and diversity. First, we keep the model exposed to varied candidate captions by sampling from the initial state before RL launched. Second, a novel criterion named max-CIDEr is proposed to serve as the reward for promoting diversity. We combine the above-mentioned offpolicy strategy with the on-policy one to moderate the exploration effect, further balancing the diversity and accuracy for human-like image captioning. Experiments show that our method locates the closest to human performance in the diversity-accuracy space, and achieves the highest Pearson correlation as 0.337 with human performance.