A01One shot learning based on computational theory of higher visual cortex
One shot learning is one of the most amazing generalization capabilities of the brain. For example, we can recognize someone’s face in any orientation after we learn only one example of that person’s face. While one shot learning has long been a challenging problem in AI research, it has not yet seen a definite algorithm. In this research plan, we aim at developing a high-performance one-shot learning algorithm by fusing neuroscientific findings in higher visual cortex and machine learning techniques including deep generative models.