Cycle Text-to-image GAN With BERT

Tsue Trevor, Sen Samir, Li Jason. Arxiv 2020

[Paper]    
Attention Mechanism BERT Model Architecture Reinforcement Learning

We explore novel approaches to the task of image generation from their respective captions, building on state-of-the-art GAN architectures. Particularly, we baseline our models with the Attention-based GANs that learn attention mappings from words to image features. To better capture the features of the descriptions, we then built a novel cyclic design that learns an inverse function to maps the image back to original caption. Additionally, we incorporated recently developed BERT pretrained word embeddings as our initial text featurizer and observe a noticeable improvement in qualitative and quantitative performance compared to the Attention GAN baseline.

Similar Work