http://feedbacknet.stanford.edu/
An alternative that can achieve the same goal is a feedback based approach in which the representation is formed in an iterative manner based on a feedback received from previous iteration’s output. We establish that a feedback based approach has several fundamental advantages over feedforward: it enables making early predictions at the query time, its output naturally conforms to a hierarchical structure in the label space (e.g. a taxonomy), and it provides a new basis for Curriculum Learning.
http://feedbacknet.stanford.edu/
http://jmlr.org/proceedings/papers/v38/lee15a.pdf Deeply Supervised Networks
https://arxiv.org/pdf/1708.04483v1.pdf Learning with Rethinking: Recurrently Improving Convolutional Neural Networks through Feedback
https://arxiv.org/abs/1607.06854 Unsupervised Learning from Continuous Video in a Scalable Predictive Recurrent Network