Futurism Part II: Artificial Intelligence
by Ralph Broadwater, MD, CFP®, AIF®
In a previous newsletter I wrote about the exponential changes that are occurring in our modern world. I would like to spend a little time discussing one of the most exciting areas of growth – Artificial Intelligence (AI).
Artificial intelligence holds the promise of disrupting every aspect of our lives, and improving society. You may not be aware of it, but advanced AI runs behind the scenes in many of the things you use daily. Google Search uses AI to generate the best search from the universe of internet data. Amazon uses artificial intelligence to suggest other products to consumers in its web store. Netflix created an AI interface to recommend new content to subscribers. Eighty percent of all content viewed on Netflix is now due to suggestions made from its AI algorithm. Facebook, Twitter, Apple Siri and Amazon Alexa all use AI to improve user experience.
Some of the most important research and development in AI is happening at Google, Amazon, IBM, Microsoft, and AliBaba. These companies have all dedicated significant budgets to advancing AI in a diverse set of disciplines.
Google’s AI division, DeepMind, gained notoriety when its AI computer program, AlphaGo defeated the best Go player in the world, Ke Jie 3-0. This human-programmed AI dominated the “Michael Jordan” of GO. In April 2017, the scientists from DeepMind published a ground-breaking paper in Nature1, Mastering the game of Go without human knowledge. In this paper, they described the development of their next iteration of AI, AlphaGo Zero. This new program taught itself and improved its play without human involvement. By using self-play and reinforcement learning, it became its own teacher. They reported that it won 100 games to none when playing their best-in-the-world player: AlphaGo.
In late December, DeepMind reported that AlphaZero taught itself chess in only four hours after being given a few basic rules. After only four hours of self-training it beat the current world champion chess playing program, StockFish.2 Truly remarkable.
Artificial Intelligence is making significant advances in medicine. Researchers at Stanford University developed an AI image recognition program that is as accurate in diagnosing skin cancer as trained dermatologists.3 It is expected that this will be developed as an app available for smartphones to help the lay public identify lesions of concern.
IBM’s Watson Oncology is using AI at Memorial Sloan Kettering Cancer Center to help oncologists choose the best treatment option for cancer patients. Analyzing the patient’s chart data and the universe of clinical data they hope to select the best treatment for a given patient4.
Artificial Intelligence shows much promise in radiology. At the 2017 RSNA, (the radiology physicians’ largest conference) there were thirty AI companies exhibiting at the convention.
Studies have shown effectiveness in mammography, CT lung and brain cancer diagnosis, and cardiac and vascular disease diagnosis.
Just last week the Chinese “Amazon”, Alibaba, reported that its AI program was successful in outscoring humans in a reading and comprehension test developed by Stanford University researchers.5 This is a significant achievement; machine reading and comprehension have been lagging behind in other advances until now.
Some of the most important uses of AI are in the areas of robotics and autonomous driving- subjects I will write about in the next newsletter.
- Silver D, et. al. Mastering the game of Go without human knowledge. Nature 550, 354-359, 2017.
- Silver, et. al. Mastering Chess and Shogi by self-play with a general reinforcement learning algorithm. org. arXiv:1712.01815v1, Dec 2017.
- Esteva A, et. al. Dermatologist-level classification of skin cancer with deep neural networks. Nature 542, 115-118, 2017.