I think that artificial intelligence machine learning / deep learning is not a program but rather a lot of data to create a model after training.
Steps to Create a Good Deep Learning Model
1) To acquire as much data as possible and to optimize and transform it into data for learning this data
2) Hyper parameter work for Loss function / Optimizer, Layer, node, Drop out etc.
3) Checking learning time and overfit through input network control
4) Overfit verification and verification of generated model and result visaulize with pyplot
2) Develop an automation program (Python) to automatically work with 3) 4)
5) Consider ensemble learning
Since 2017, I have developed the following services by learning and applying machine learning.
(All deployed under AWS / Azure / Google Cloud service using Google Tensorflow / Keras.)
- aiMu (easy and simple AI piano music composition app)
- miseGo (App developed by React-Native. AI Module developed by Google TensorFlow)
- eXGo (App developed by React-Native)
I have been working on artificial intelligence development for two years. I think that it is my strength to have made artificial intelligence model while considering the application of artificial intelligence model to actual service.
'즐겁게 살자 > 인공지능(Google)' 카테고리의 다른 글
|What is Machine Learning / Deep Learning ? (My opinion) (0)||2019.06.07|
|내가 생각하는 딥러닝/머신러닝 개념 정리 kor (0)||2019.06.07|
|美청소년 95%가 스마트폰 사용, 인기 서비스는 유튜브 (0)||2018.06.29|
|[알아봅시다] 공중전도 무인화 대세 .. 세계 각국 개발 경쟁 치열 (0)||2018.06.21|
|음... 삼성 AI ... (0)||2018.06.20|
|인공지능 로봇 쉐프 레스토랑 (0)||2018.06.18|