Understanding the Neurons in Neural Networks (Part 1): Artificial Neurons

摘要

Artificial intelligence (AI) and machine learning (ML) are two of the industry's hottest topics. Thanks to successes covered in the press, with AI beating the world's best Go players [1], and failures, such as autonomous vehicle accidents [2], AI has become part of the lingua franca. While AI and ML have been democratized thanks to cloud-based tools such as TensorFlow, these huge powerful platforms can seem intangible if you are trying to understand how ML works "under the hood." In this series, we go back to basics and explore most of these systems' basic building block, the neural network. Along the way, there are plenty of examples to try, examples of cool and obscure ML projects, and at the end, we'll even bestow an Arduino with a brain. AI and ML are "the" computing challenges of our time. AI, which focuses on using computers to mimic human intelligence, and ML, which targets pattern recognition of structured and semi-structured data, command significant annual investment from research projects to the development of semiconductor technology and computing platforms. And, thanks to "the cloud," the technology is easily accessible to those wanting to explore and test their ideas.