What is AI?

“The development of full artificial intelligence could spell the end of the human race” warned world-renowned physicist Stephen Hawking in 2014. Elon Musk likened improving artificial intelligence to “summoning the demon” and called it the human race’s biggest existential threat. On the other hand, Sundar Pichai Google CEO said Google is evolving from Search first to mobile first and to AI first world. Robin Li the CEO OF BAIDU said “AI will continue to be an unceasing core for Baidu’s innovation”. Recently, Google’s AlphaGo defeated Lee Sedol. The machine’s sweeping victories have once again made AI a hot topic. The purpose of this blog is to provide an overview of AI.

What is Artificial intelligence?

AI /ML is an approach to computing which is different from traditional programming. In traditional programming, engineer explicitly writes the step by step instructions for computer to follow while in AI programmers don’t encode computers with interaction. They just train them. If you want a machine to recognize cat, you don’t tell it to look for whiskers, eyes, ear and fur. You simply show it a picture of cat and then one more and then one more and then thousands and eventually the computer works things out. We are not giving any characteristics of the data. We are just providing raw data and computer figure out the rest

Why is it so popular now? What happened?

AI is not a new approach. It has been around for decades? Genesis of AI started in summer of 1956 in Dartmouth when John McCarthy and group of four other professors came together and said “Hey look computers are sophisticated, people are sophisticated and we ought to be able to teach computers these tasks that people can do. This was an ambitious goal in the summer of 1956 when we only had mainframes with punch cards. Let me give you an example, its mid 1950’s, height of cold war and US want to translate desperately from English to Russian. Researchers took 60 Russian sentences and used AI to convert them to English sentences and it worked flawlessly in this constrained environment. However, when the same model was applied to real world it failed miserably. Let me give you an example sentence “The spirit is willing but the flesh is weak” was converted to “whisky is strong but the meat is rotten”. It’s obvious why the translation happened but at the same times it’s easy to understand why it is wrong It was not able to capture the meaning or the semantics but it was more of a syntactical conversion So if you ask a question gee, why are people so excited about this AI now when it have been out since 1950’s the answer scale and data , due to the availability of the massively distributed computational systems and large swaths of data.We have been able to throw lot of machines and lot of data at solving the problem. Rather than that one mainframe in 1950’s we now have these millions of computers that more powerful than mainframes solving this problem.

How is it being used now?

Facebook uses it to determine which stories should show up in our news feed. Google photo uses it to identify faces of our friends. Self-driving cars use machine learning to avoid accidents .Netflix is using it to recommend movies and episodes. It is being incorporated into medicine to help doctors detect cancer and save lives, and predict weather.

Will AI destroy us?

Will in future a robot write the perfect blog post about AI?

Not anytime soon, say computer scientists. Despite their fast advances, robots remain too woefully single-minded to give humanity a giant pink slip. Teaching computers to read and then answer questions about that content is still a work in progress. For example, when scientist had computers read eighth-grade science texts, they could answer only about 60 percent of the questions on a test. “Understanding a single sentence can be a lot more complicated than playing chess or Go. Computers, have yet to demonstrate that they “can solve fuzzier problems where things are more nuanced.”

There you have it. Your job is safe for now.

After the final game, Lee Sedol said that he had already learnt from AlphaGo. I thought this was insightful. AlphaGo was taught by Humans to become more intelligent and now here it was teaching a Human to become better version of themselves.