Hi Folks! 2018 is coming to an end. A year passed by just like a moment. It's time to do some math as to what we achieved in what could be termed as the fastest year for many people including me.
AI became the most hyped topic of discussion among the Tech giants in 2018. So, since July I have been waiting for what could be termed as a major breakthrough. Because 2016 and 2017 were such breakthrough years with AlphaGo, AlphaZero and deeplearning gaining huge momentum. But compared to the previous two years, in 2018 no major breakthrough has happened so far. Atleast among the frontrunners in the field. Because we have no way of knowing in specific details as to what kind of AI the defense departments of US, Russia and China are building.
If I make a kind of a sweeping statement as no major breakthrough, it becomes necessary to define what I call as breakthrough. Anything that drives a shift in conscience among the people working in the field or a new tecnhology that impacts a whole lot of other related technologies is a (major)breakthrough. Like IBM deepblue in 1997 or AlphaGo in 2016.
But some significant advancements have been made in 2018 which are worth-knowing and worth-cherishing. The following are my personal favourites and am sure you will like it too:
Self Driving Cars
In July 2018, Waymo,the secretive subsidiary of Google’s parent company, Alphabet Inc., announced that its test vehicles had traveled in automated mode for over 8,000,000 miles (13,000,000 km), increasing by 1,000,000 miles (1,600,000 kilometres) per month. Waymo has been working on self-driving technology for nearly a decade. They began as the Google self-driving car project in 2009. Presently they call themselves an independent self-driving technology company with a mission to build the most experienced and reliable driver that you can trust just like that.
I heard that they are planning to launch the world’s first commercial driverless car service in early December. It will operate under a new brand and compete directly with Uber and Lyft.
In August, Deepmind announced the results of the first phase of their joint research partnership with Moorfields Eye Hospital, which could potentially transform the management of sight-threatening eye disease.
The results, published online in Nature Medicine, show that their AI system can quickly interpret eye scans from routine clinical practice with unprecedented accuracy. It can correctly recommend how patients should be referred for treatment for over 50 sight-threatening eye diseases as accurately as world-leading expert doctors.
Since November, they are working on the next research challenge to help clinicians predict eye diseases before symptoms set in.
In June 25, OpenAI Five, a team of five neural networks, defeated amateur human teams at Dota 2. At that time they played with restrictions. Soon after they set a goal to beat a team of top professionals at The International (Dota 2) in August.They were initially not confident that they would win because Dota 2 is one of the most popular and complex esports games in the world, with creative and motivated professionals who train year-round to earn part of Dota’s annual $40M prize pool (the largest of any esports game).
But on August 5, OpenAI Five won a best-of-three against a team of 99.95th percentile Dota players: Blitz, Cap, Fogged, Merlini, MoonMeander and — four of whom have played Dota professionally — in front of a live audience and 100,000 concurrent livestream viewers. The human team won game three after the audience adversarially selected Five’s heroes. They also showed their preliminary work to introspect Five’s view of the game, including its probability of winning, which made predictions surprising to the human observers. These results show that Five is a step towards advanced AI systems which can handle the complexity and uncertainty of the real world.
Have a look at the live broadcast - it is fun:
In 2018, OpenAI launched the Debate Game, which teaches machines to debate toy problems in front of a human judge. In this game, One debater tells the truth, the other lies. An image is visible to the two debaters, and they can draw rectangles on the image and talk to the judge. The judge sees only the rectangles: the image is hidden. Each debater is also allowed to reveal a single pixel to the judge. The pixel must be chosen carefully, since they can only reveal one pixel total over the whole debate.
The purpose of this work is to research whether such an approach may assist in auditing AI decisions and in developing explainable AI.
Facebook AI research
Facebook hasn't released anything super-impressive in the field of AI till date. However, they keep on publishing their research works at FAIR. Today, deepmind pose a serious challenge to facebook ai. First, they lost on a bet to acquire deepmind before Google. In the AI race, they are left far behind by deepmind and google. I say that because recently they tried to reproduce deepmind's AlphaGoZero and is claiming to have beaten defeated world champion professional Go players, but we don't know the names. The good thing is that they have opensourced their ELF OpenGo, an AI bot based on their existing ELF platform for Reinforcement Learning Research.
That's all for today. If hope you enjoyed reading this post. If you find this informative, do share it with your friends...