Abstract: There has been a gap between artificial intelligence and human intelligence. In this paper, we identify three key elements forming human intelligence, and suggest that combines these elements and is thus a way to bridge the gap. Prior researches in artificial intelligence either specify abstraction by human experts, or take abstraction as a qualitative explanation for the model. This paper aims to learn abstraction directly. We tackle three main challenges: representation, objective , and algorithm. Specifically, we propose a partition structure that contains pre-allocated abstraction neurons; we formulate abstraction as a constrained optimization problem, which integrates abstraction properties; we develop a network evolution algorithm to solve this problem. This complete framework is named ONE (Optimization via Network Evolution). In our experiments on MNIST, ONE shows elementary human- like intelligence, including low consumption, knowledge sharing, and lifelong .

Source link
thanks you RSS link
( https://www.reddit.com/r//comments/9fa35v/r__abstraction_learning/)


Please enter your comment!
Please enter your name here