Study on a New Cognitive Algorithm to Improve Traditional Algorithms

Abstract

Traditional instruction-based and automatic computations are passive, dominating, and target-driven computing systems with limited dependence on heuristic procedural information. In contrast, our new cognitive computing is a computational system that implements human natural intelligence, such as reasoning and learning, aiming to maximize the cognitive power of machines while minimizing resources. The original goal of Artificial Intelligence (AI) to simulate human intelligence is not being achieved at an appropriate level, and this paper aims to analyze human brain activity in depth and to define the need for human-like cognitive algorithms based on the knowledge experienced in real life, and propose a new cognitive algorithm, such as deep learning, rather than to improve traditional methods. Therefore, the paper proposes a new cognitive algorithm that can revolutionize existing algorithms by implementing information processing methods in human brain; learning data based on experience, generating and evaluating conflicting hypotheses, providing results, discovering data patterns, mimicking the process or structure of natural learning systems, extracting meaning from textualized data using Natural Language Processing (NLP), and extracting features from static images, animation, speech, and sensors using deep learning.