Google's artificial intelligence (AI) ethics board, established when Google acquired London AI startup DeepMind in 2014, remains one of the biggest mysteries in tech, with both Google and DeepMind refusing to reveal who sits on it. Google set up the board at DeepMind's request after the cofounders of the 400 million research-intensive AI lab said they would only agree to the acquisition if Google promised to look into the ethics of the technology it was buying into. Business Insider asked Google once again who is on its AI ethics board and what they do but it declined to comment. A number of AI experts told Business Insider that it's important to have an open debate about the ethics of AI given the potential impact it's going to have on all of our lives. Artificial intelligence is the field of building computer systems that understand and learn from observations without the need to be explicitly programmed, as defined by Nathan Benaich, an AI investor at venture capital firm Playfair Capital.

Over at Aeon, Caleb Scharf, director of Columbia's Astrobiology Center, has written an essay contemplating the substrate of minds. Technological progress has spurred many futurists to predict a singularity, a moment when machine intelligence, or AI, overtakes humanity (see also the recent development of our Go overlord). This line of thought has found its way into astrobiology (Dr. When we look for extraterrestrial life, several people have proposed that it might be worthwhile to look for machines rather than the typical little green men. The squishy nature of most biological life does not necessarily lend itself well to interstellar travel.

First went checkers, then fell chess. Now, a computer program has defeated the world's top player in the ancient east Asian board game of Go -- a major milestone for artificial intelligence that brings to a close the era of board games as benchmarks in computing. At the Four Seasons Hotel in Seoul, Google DeepMind's AlphaGo capped a 3-0 week on Saturday against Lee Sedol, a giant of the game. Lee and AlphaGo were to play again Sunday and Tuesday, but with AlphaGo having already clinched victory in the five-game match, the results are in and history has been made. It was a feat that experts had thought was still years away.

The competition will take a while. It doesn't start until 2017, and won't pick a winner until early 2020. DARPA will even have to create a giant wireless testbed to see how the competitors fare in relatively realistic conditions. It could be worthwhile, though, as the winner will scoop up a 2 million prize. The institution notes that there could be clear advantages to AI-based radios in the military, which could keep communications up and running on the battlefield.

Teams representing the University of California, Berkeley and Arizona State compete in the Grand Final of last year's Heroes of the Dorm tournament at the Shrine Expo Hall in Los Angeles. By the end of this week, four teams will advance to determine a champion after an intense multi-week competition. Their names will look familiar -- UConn, Miami and Oregon -- and monetary stakes are high, to the tune of 500,000. They're battling with video game controllers in the growing arena of competitive video gaming, whose increasing popularity has attracted the attention of big names in tech and media, from Electronic Arts to ESPN to Yahoo, as they see a growing audience eager to learn about the competitive scene and engage more in games at the center of its rise. 'We have the early markers of what will ultimately make eSports mainstream," says Joost van Dreunen, CEO of SuperData Research, which gathers data on the global games market. But it could require a generational shift before competitive video gaming -- known to many as "eSports" -- formally becomes mainstream entertainment. On April 3, the "Heroic Four" will be determined in Heroes of the Dorm, a competitive video game tournament hosted by Blizzard Entertainment, based on its action game Heroes of the Storm. For the second year, teams representing colleges from across the U.S., including the University of Connecticut and Arizona State University, are playing for glory and more than 500,000 in scholarships and prizes, including a free ride through school for the winning team. Fans watch the action online on ESPN, Twitch and YouTube, and they can even join tournament pools, where the winner with the most accurate bracket snags 10,000. It's the latest example of competitive video gaming's increased following, as younger fans gravitate towardeSports. The market is valued at 747 million, according to SuperData, and is expected to more than double to 1.9 billion in three years. The rising audience -- SuperData estimates it at 134 million as of last year -- is pushing video game publishers and cable networks to create competitive video game experiences and explore broadcasting options. The eSports market is young. Whalen Rozelle, Director of eSports at Riot Games -- makers of the hit competitive game League of Legends -- says it's still in "our pre-teen phase," with plenty of room to grow. "The industry still hasn't really figured out'is every game an eSport?

The nonnegative matrix factorization (NMF) is widely used in signal and image processing, including bio-informatics, blind source separation and hyperspectral image analysis in remote sensing. A great challenge arises when dealing with a nonlinear formulation of the NMF. Within the framework of kernel machines, the models suggested in the literature do not allow the representation of the factorization matrices, which is a fallout of the curse of the pre-image. In this paper, we propose a novel kernel-based model for the NMF that does not suffer from the pre-image problem, by investigating the estimation of the factorization matrices directly in the input space. For different kernel functions, we describe two schemes for iterative algorithms: an additive update rule based on a gradient descent scheme and a multiplicative update rule in the same spirit as in the Lee and Seung algorithm. Within the proposed framework, we develop several extensions to incorporate constraints, including sparseness, smoothness, and spatial regularization with a total-variation-like penalty. The effectiveness of the proposed method is demonstrated with the problem of unmixing hyperspectral images, using well-known real images and results with state-of-the-art techniques.

As I walked around my office on a recent morning, a female voice on my iPhone narrated the objects I passed. "Brick," "wall," "telephone," she said matter-of-factly. The voice paused when I came upon a bike hung on a wall-mounted rack, then intoned, "bicycle." The voice is part of a free image-recognition app called Aipoly that's trying to make it easier for those with vision impairments to recognize their surroundings. To use it, you point the phone's rear camera at whatever you want it to identify, and Aipoly will speak what it sees (or, at least, what it thinks it sees) and show the object's name on the phone's display.

Scheduled Sampling was at NIPS last year; the reviews are also online. This is actually the third time I've tried to make my way through this paper, and to force myself to not give up again, I'm writing my impressions here. Given that this paper is about two things I know a fair amount about (imitation learning and neural networks), I kept getting frustrated at how hard it was for me to actually understand what was going on and how it related to things we've known for a long time. So this post is trying to ease entry for anyone else in that position. What is the problem this paper is trying to solve?

Nearly everything at Google has an acronym. Machine learning, the artificial intelligence method for processing reams of data, currently all the rage across Google, is just "ML" inside the company. On Wednesday, Google presented its newly assertive push for the enterprise, hosting its inaugural cloud developer conference in San Francisco. In fact, Google unveiled a handful of new offerings that, in essence, pour ML all over the cloud. "I've become convinced that there's a new architecture emerging," an exuberant Eric Schmidt, chairman of Google parent Alphabet, said from the stage.

I hope you've followed my previous articles on ensemble modeling. In this article, I'll share a crucial trick helpful to build models using ensemble learning. 'How to choose the right models for your ensemble process?' Imagine the following scenario (great, if you can relate to it). You are working on a classification problem and have built 1000 machine learning models. Each of the model gives you an AUC in the range of 0.7 to 0.75 .