초록

용어보기논문에서 용어와 풀이말을 자동 추출한 결과로, 시범 서비스 중입니다. AI-Helper

Abstract ▼ AI-Helper

Recently, neural networks attracted public attention as a new computational paradigm, which makes an attempt to solve many difficult problems with massive interconnections of simple processors. The slowness of learning, however, is the major abstacle for its application to real world problems. So i...

Recently, neural networks attracted public attention as a new computational paradigm, which makes an attempt to solve many difficult problems with massive interconnections of simple processors. The slowness of learning, however, is the major abstacle for its application to real world problems. So is required the systematic analysis of widely used learning algorithms such as backpropagation, and faster learning methods. This paper classifies the conventional acceleration techniques of backpropagation learning into three categories, heuristic based, numerical method based, and learning strategy based techniques, and has an attempt to compare them through their time and storage complexity. Based on this comparative classification, presented are some discussions about points to be considered when faster learning techniques are proposed.