CyCADA: Cycle-Consistent Adversarial Domain Adaptation
This addresses the problem of visual recognition in unseen environments for applications like autonomous driving, with incremental improvements in combining existing techniques.
The paper tackles domain adaptation by proposing CyCADA, a model that adapts representations at both pixel and feature levels using cycle-consistency and adversarial training without aligned pairs, achieving new state-of-the-art results in tasks like digit classification and semantic segmentation from synthetic to real domains.
Domain adaptation is critical for success in new, unseen environments. Adversarial adaptation models applied in feature spaces discover domain invariant representations, but are difficult to visualize and sometimes fail to capture pixel-level and low-level domain shifts. Recent work has shown that generative adversarial networks combined with cycle-consistency constraints are surprisingly effective at mapping images between domains, even without the use of aligned image pairs. We propose a novel discriminatively-trained Cycle-Consistent Adversarial Domain Adaptation model. CyCADA adapts representations at both the pixel-level and feature-level, enforces cycle-consistency while leveraging a task loss, and does not require aligned pairs. Our model can be applied in a variety of visual recognition and prediction settings. We show new state-of-the-art results across multiple adaptation tasks, including digit classification and semantic segmentation of road scenes demonstrating transfer from synthetic to real world domains.