Streaming Self-Training via Domain-Agnostic Unlabeled Images

Abstract

We present streaming self-training (SST) that aims to democratize the process of learning visual recognition models such that a non-expert user can define a new task depending on their needs via a few labeled examples and minimal domain knowledge. Key to SST are two crucial observations: (1) domain-agnostic unlabeled images enable us to learn better models with a few labeled examples without any additional knowledge or supervision; and (2) learning is a continuous process and can be done by constructing a schedule of learning updates that iterates between pretraining on novel segments of the streams of unlabeled data, and fine-tuning on the small and fixed labeled dataset. This allows SST to overcome the need for a large number of domain-specific labeled and unlabeled examples, exorbitant computational resources, and domain/task-specific knowledge. In this setting, classical semi-supervised approaches require a large amount of domain-specific labeled and unlabeled examples, immense resources to process data, and expert knowledge of a particular task. Due to these reasons, semi-supervised learning has been restricted to a few places that can house required computational and human resources. In this work, we overcome these challenges and demonstrate our findings for a wide range of visual recognition tasks including fine-grained image classification, surface normal estimation, and semantic segmentation. We also demonstrate our findings for diverse domains including medical, satellite, and agricultural imagery, where there does not exist a large amount of labeled or unlabeled data.

SelfD: Self-Learning Large-Scale Driving Policies From the Web

“…Consequently, our approach goes beyond prior policy distillation approaches to handle scenarios where supervision by the teacher model may be potentially noisy and unsafe. We also note the relationship between such distillation and semi-supervised training via pseudo-labeling [8,40,45,62,64]. However, as far as we are aware, we are the first to develop a pseudo-labeling based self-training method for learning safe driving policies from complex scenes with diverse navigation data, camera perspectives, geographical locations, and weathers.…”