Transfer Optimistic Outcome-based Learning for Mature Behavior of Machine in Deep Learning

摘要

This paper focuses on experiences in every experiment in an optimistic manner as an input to the machine to train and create a Deep learning mechanism with previous optimistic outcome-based learning based on an evaluation by Bayes Theorem instead of hierarchical representation only. Here, we want to suggest that it is not favorable to depend on data only. Instead of focusing on data only to train the machine every time past experiences must be counted as outcomes. These outcomes, further transfer to the machine along with new data can change the approach of a machine to learn, and especially in Deep learning to train the model will be more affirmative and its hierarchical representation gains a sense of previous experiences. This paper focuses on experiences in every experiment in an optimistic manner as an input to a machine to train and create a Deep learning mechanism instead of hierarchical representation only. Also, these experiences must be optimistic in sense of often realistic, linear and high dimension. Knowledge-based on such optimistic experiences has a scientific value. We can use the Bayes formula repeatedly to increase correctness.