Today, Google posted a blog on their AI advancements. The article talked about advances in natural language processing, computer vision and more. They also released a video discussing some of their latest developments with Google Assistant.
The tech giant has been adding more to Google Assistant. The assistant can now schedule meetings for you as well as make restaurant reservations for you by calling the restaurant.
Google is really pushing the limits with their AI technology and it will be interesting to see how they continue to advance in the coming years.
A few years ago, I read an article on the top 12 Google interview questions. As a developer, I’ve never had the pleasure of interviewing at Google, but as I read through the questions, I was impressed by how much they required understanding of not just language syntax and frameworks, but also larger computer science topics such as memory management and algorithms. One question that stuck out was “How would you move Mount Fuji?”, which requires problem solving completely outside of traditional programming knowledge.
A few months ago, I read another article on Google’s quest to teach its computers to be creative. The article mentioned one experiment that involved showing a computer thousands of images of cats and asking it to understand what makes an image a “cat” image. The computer easily mastered that task, but when asked to produce its own image of a cat, it showed only gray noise. The experiment highlighted an important problem with AI research: computers excel at recognizing patterns in large quantities of data, but they are terrible at creating new patterns or even understanding what the purpose of those patterns is.
At the same time, I recently played a game called Rise of the Triad for my Amiga 500 emulator on my phone. For those who haven’t played it, Rise
Google recently held its third annual coding competition. Contestants were asked to write a program that plays the game setting up servers. The game involves setting up servers and connecting them together with cables. Each server can only be connected to two other servers, and each cable has a cost between 0 and 7 that must be paid. The goal is to connect all the servers while spending as little money as possible.
The contest was won by a student in Germany who figured out how to solve the problem in 1/30 of a second, or about 3 million times faster than the next fastest competitor, who had already won last year’s contest. It’s hard to imagine what kind of algorithm could do something like that. But perhaps more interesting than this algorithm’s staggering efficiency is how it came about: it was written by an artificial intelligence program called Mogul.
To understand how Mogul works, let’s first talk about how people try to solve such problems, starting with Google’s contest from last year, which asked contestants to write a program that plays the game of setting up servers. Last year’s winner was an algorithm called AlphaGo Zero, which was able to play Go more effectively than any human player ever has by playing games against itself over and over again, using only what it
Google Code Jam to I/O for Women is a Google-sponsored contest for women in computer science. This year, the contest takes place on May 9th and 10th, and registration is open now!
Google Code Jam to I/O for Women is an opportunity to challenge yourself as a software engineer. It’s also a chance to meet new people from around the world, sharing your experience of coding. Each year, thousands of people participate in Code Jam around the world – and many of them say it’s one of their most fun coding experiences of the year.
This year, we’ve changed up the format to make it even more accessible! We’re offering two tracks: Beginner and Advanced. The Beginner track will be an introduction to coding challenges, while the Advanced track will provide greater challenges that are appropriate for those with previous coding contest experience. Both tracks will be available in English, Spanish and Portuguese.
We’ll also be adding a new feature this year: Google Hangouts On Air! These will provide an opportunity to ask questions of other contestants using Google+ Hangouts while watching live streams on YouTube.
Regardless of your skill level or language preference, we’d love for you to join us this May!
Google announced a coding competition on Kaggle to reward the best developers and machine learning experts in a challenge of image recognition. The contest will award $1 million to the winning team, which will be divided among the top five finishers. The competition is called “The Nature Conservancy Fisheries Monitoring” and it was designed with help from marine biologists to help solve an important problem facing nature conservation: how do you monitor fishing boats?
Image recognition is an important part of artificial intelligence, which is going to change the way we live and work. While machines aren’t going to replace humans anytime soon, they’re getting better at doing certain tasks that are too boring or repetitive for us, like monitoring for illegal fishing.
The competition requires participants to create a computer program that can identify fish species from photos taken in the wild by cameras placed on fishing vessels. The program has to look at an image and determine if one of seven species of fish is present. The goal is to get data that can help governments and environmental organizations stop illegal fishing and overfishing, which have contributed to the decline of many key fish species around the world.
When you have a coding competition at Google, it’s not your run-of-the-mill hackathon. There are tasty snacks and benefits, like massages, to help the 10,000+ students in attendance destress.
The company has hosted the global competition for over a decade and they’re now extending it to Europe for the first time. This year’s event takes place across five cities (Barcelona, Berlin, London, Warsaw and Zurich) on 4 November 2017.
“We are looking for creative coders who are able to solve complex problems,” says Kasia Chmielinski, University Relations Manager at Google. “We want people who think beyond the brief.”
There is no specific criteria to enter the competition. It is open to anyone who is 18 years old or over by 4 November 2017. The only thing you need is a basic understanding of programming. Students from around the world can also participate online.
The competition will push you to your limits and test your ability to think quickly and logically under pressure. It is not about how quickly you can write lines of code, but how well you can solve problems with them. That said, you will be given a short time limit – an hour – to complete each challenge.
During a recent coding competition hosted by Google, 4 teams of computer programmers competed to see which team could build the best artificial intelligence (AI) program. The challenge: create an AI that can play StarCraft II, a video game where players compete in a futuristic sci-fi universe. While this may sound like a fun competition, it’s actually an important step toward creating true artificial intelligence, a goal for which we are still far from reaching.
In order to understand why this competition was so significant, it’s important to note what makes an AI “smart”. In the past, artificial intelligence programs have been able to beat expert chess players and do complex mathematical calculations. However, these programs only do one thing well because they rely on brute force tactics–simply trying many possibilities until finding the right answer–and can’t learn how to improve on their own.
This is not true intelligence because it is not sustainable in the long run; eventually, one of those brute force calculations will take too long or become too expensive. For example, how would you make an AI that can do your taxes? It simply isn’t possible because there are too many variables involved and not enough information available for the program to go through every single