As a little project, I'm working on building out a neural network that learns to play a snake game. The game is composed of a max 20x20 board where each cell is either an apple, a snake, or a wall. One way I thought of encoding that data in the input layer would be to have a 20x20 400 neuron input layer where each neuron has one of 3 different states signaled by neuron value. Say, 0 apple, 0.5 snake, 1 wall. Is encoding multiple states in a single neuron like this a good idea?

It all depends on your skills as a developer. If you do know how to work with many threads on many cores, I'd go for a cheap xeon/amd server with 2 or 4 cpu sockets to get up to 64 cores, a min of 1 gigabyte of ram per core, A BOOTABLE RAM DISK ON PCI-EXPRESS WITH AUTOMATIC BACKUP ( SSDs are ridiculous and overrated, they burn out so easily and they're not worth the risk for long-term storage purposes) and a fast HD (10k rpm minimum) as storage. For the GPU, honestly, unless you plan on working with CUDA/ opencl, anything is fine because you'd rarely compute on it. But if you will develop GPU-"powered" neural networks and if wattage isn't of a concern for you, given a proper thermal dissipation, there are many AMDs that can pack a punch for little money both in single and double precision. But if you don't know how to take advantage of multithreading and if frameworks are what you have in mind, whatever you buy, as long as it is fast, "it's gonna be fine".

Flippy, the burger flipping robot only lasted two days before the plug was pulled, temporarily, suggesting that the future we dream about is going to take a whole lot longer to arrive. But it hit with a thud as reality quickly sunk in. The growing robotic list goes way beyond factory production lines to include self-driving cars, delivery drones and especially the one we encountered this past week, the burger-flipper. On its surface, the idea sounds like a no-brainer: Replace young men and women who can't take the heat and monotony of standing over a hot griddle -- and end up quitting in weeks, if not months -- with a robot. A $100,000 robot can grill hamburgers to perfection all day long.

Dulce Sloan finds out that the downside (and upside) of facial recognition software is that people of color often go undetected.

Hollywood movies telling us since long those Robots are coming to get us, people were enjoying these indications as jokes and science fiction movies. Some time back New York Times and Mr. Musk told us the same thing but this time we are serious and anxious. In early days if some one answers to our very old-style question "What do you do" regardless of your origin or culture or from you come from or language you speak we get same question very often. If the answer was "I do artificial intelligence" and immediate reaction was "Why don't you have natural one". Today if the answer to the same question is "I do artificial intelligence"; immediate reaction is "Please don't hurt me / us".

Artificial intelligence was once a phenomenon restricted to sci-fi movies, but technology has finally caught up with imagination. Now AI has become reality and amazingly, most people encounter some form of artificial intelligence in their everyday lives. For example, AI has dramatically improved the technology in our homes. Amazon's Alexa can carry out a number of functions via voice interactions, such as playing music, answering factual questions and even making online purchases. And now, with the recent launch of AmazonGo, AI allows shoppers to simply walk out of a shop with their purchases and automatically get charged, sending an invoice directly to their phone.

Gervais' beaked whales are easily one of the most elusive mammals to swim through our oceans. Most of the information we have about them comes from studies of corpses that have washed ashore, and the first live whale was only spotted about 20 years ago. On February 27, photographer and videographer Patrick Dykstra captured what may be the first drone or aerial footage of Gervais' beaked whales. He was filming about three miles off the west coast of Dominica in the Caribbean Sea. Dykstra and his Picture Adventure Expeditions team accidentally came across the rare beaked whales when they were filming sperm whales for an upcoming production.

Apple is reportedly coming out with a cheaper version of its $349 HomePod, Business Insider reports. The company is following in the footsteps of Google and Amazon, both of which released'discount' versions of their own smart speakers, making the technology more accessible to a wider audience. The cheaper HomePod would cost somewhere between $150 and $200, and it would be smaller. The new report says it could be available as soon as this fall. Apple might be releasing a cheaper, smaller version of its HomePod speaker.

When we're asked, "What is artificial intelligence?" Perhaps it's a sassy-talking technology like Siri from Apple, or helpful humanoid counterparts like those depicted in The Jetsons. Some might even imagine sophisticated robots threatening to extinguish the human race. Nowadays, there are as many definitions of AI as there are companies trying to pitch AI solutions. So, how do law firms know how to incorporate artificial intelligence? When is the right time to use legal AI? Think of AI as computers performing tasks previously thought to require human intelligence.

If you love Andrew Ng's first Coursera course on machine learning as much as I do, you were equally hyped when you heard that deeplearning.ai Since everybody's on a tight schedule, let's try the impossible and finish a course that is laid out to last one month in one week. Let's not rush through though, but actually understand the material. And of course, we'll do it while continuing our 40h/week job. What are the advantages of finishing the course quickly you ask?