Daily Archives: July 11, 2017

Google's AI program, AlphaGo, went up against -- and defeated -- Chinese Go champion Ke Jie (on the left) at the Future of Go Summit in May in China. The match took place a year after AlphaGo bested Lee Sedol, world number two Go player.

AlphaGo may have defeated humans at board games, but its creators really just want us to be buddies.

In a new project named the People + AI Research Initiative (PAIR), Google's researchers are looking at the relationship between humans and artificial intelligencein the hopes of making the latter more useful to the former, the tech giant announced on its blogon Monday.

The company says it'll rethink AI on three levels: How we can use it as a tool in everyday life, how professionals in all fields can use it to make their jobs easier and how practical AI development can be taught to engineers.

Google isn't the only one making big moves to help develop the nascent field. On Monday, the Ethics and Governance of Artificial Intelligence Fund, helmed by Harvard University's Berkman Klein Center for Internet & Society and the MIT Media Lab, pledged $7.6 million to support the creation of AI that serves public interest. Plus,the tech giant last year partnered with Amazon, Facebook, IBM and Microsoft to create a new not-for-profit called the Partnership on Artificial Intelligence to Benefit People and Society.

Google says, as part of PAIR, it will introduce new open-sourced tools and educational material as well as publish research to help push AI along.

Tech Enabled: CNET chronicles tech's role in providing new kinds of accessibility.

Batteries Not Included: The CNET team reminds us why tech is cool.

Read this article:

Google's new PAIR project wants to rethink how we use AI - CNET

AI is becoming a way of life for many of us. We check on flights using a chatbot like Mezi, we benefit from the AI within the booking engine used at Hoppers website, and we are sending messages to businesses easier thanks to the machine learning at Yelp.

It should not come as a big surprise when the AI improves, advances, and becomes even more helpful. After all, taking a cue from the human brain, AI is always adapting, looking for new ways to help us on a constant iteration cycle. The engineers behind AI are keen to make the technology more powerful and integrated into our daily workflow, even when things get really complex.

Thats why several companies are not interested in spinning their wheels when it comes to AI. Today at MB 2017, four companies made a splash with announcements that are intended to make their services even more competitive and help make your life easier.

One interesting upgrade has to do with the Mezi chatbot. The app uses AI algorithms to help with flight searches and other duties but is also powered by human agents. Today, they have announced Mezi for Business. The new service, intended for travel agents and corporate travel reps, will improve efficiency and productivity.

Similar to the consumer app, it employs algorithms to help with travel booking and managementand much more.

We have decided to go all-in on travel, saysSwapnil Shinde, the CEO and founder of Mezi, speaking at MB 2017. We empower businesses with a suite of travel bots that automate requests. For travel agents we offer a state-of-the-art travel dashboard.

Another example of gaining traction Yelp is using machine learning to facilitate and improve the interactions between customers and businesses. Its fine-tuned behind the scenes by an AI. 35,000 messages are fed through their machine learning tech. They use data from service companies to find out about geo-fencing parameters. They extract data about the services as well. Yelp is also using machine learning to weed through content and verify it, making sure that five star review of an auto repair business is valid.

The last feature, requesting a quote from a business, is also AI enabled. For example, it makes sure a business matches the request.

We estimate that every month, Yelp sends billions of dollars of leads to local service businesses listed on our site through the Request A Quote feature, says Jim Blomo, the director of engineering at Yelp. Growth of this feature has been through the roof, and a lot of that progress can be attributed to the machine learning work on this product, allowing us to surface the most useful and relevant businesses when a consumer types iPhone 7 screen repair or overflowing toilet into Yelp.

Another company, GobTech, is using AI in its iOS and Android app called Neural Sandbox. The apps let you experiment with neural networks. At MB 2017, the startup is launching a way to compare neural networks called Gauntlet. Users can compare their score against other users using the Google Play leaderboard.

GobTech is exploring new frontiers in AI for gaming using a unique combination of neural networks and genetic algorithms, says Gabriel Kauffman, the CEO of GobTech. This combo, known as neuroevolution, is a way for neural networks to evolve through natural selection, in our case to learn to play a game by itself.

Meanwhile, Hopper is using machine learning to improve its back-end booking agent. Its an effort to make booking work more like you have a human helping you find the best travel deals. Maggie Moran, the Head of Product at Hopper, explained how the AI bunny empowers travelers about how to find the best deals.

GoPro revealed how they are using AI.Meghan Laffey, the VP pf Product at GoPro, explained how the app is central to their product offering. The phone has made it easy to go from capturing to sharing, she says. Its been a challenge to go from the experience to the actual playback.

A new feature called Quik Stories allows users to film and edit videos without the hassle of watching all of your footage. With a single tap, stories are generated automatically. Algorithms analyze content and find the best moments, syncing them to music.

These announcements show how AI will ultimately gain traction by iterating, improving, and capturing new audiences.

The ability to use AI within an app is nothing new. What will create a differentiator in the long run is when companies keep enhancing the AI, when the machine learning power an app or website is so compelling that it attracts new users.

Read more from the original source:

5 AI-powered companies gaining traction for 2017 - VentureBeat

AI Ventures will direct its investments towards AI, robotics, autonomous vehicles and data and cloud technology. Along with funding, it will also offer companies it invests in both mentorship and support at its Silicon Valley headquarters. "One of the biggest challenges entrepreneurs face is knowing if they're building the right product for the right market. We can help them navigate that uncertainty, and we're committed to doing so in a founder-friendly way because their success is our success," said TRI VP Jim Adler in a statement. AI Ventures says it will also be proactive in how it tracks down companies to invest in. Rather than only considering pitches from those seeking investors, it will also seek out and support new companies aiming to solve key research challenges the fund is interested in.

So far, the fund has invested in three startups. Silicon Valley-based Nauto designs systems for companies that monitor their drivers and road environments in order to prevent accidents and curtail bad driving. SLAMcore is a UK company that develops algorithms for smart tech, like drones and self-driving vehicles, that allow them to create a map of their surroundings and position themselves within it. And the third company, Intuition Robotics, is a social companion technology startup located in Israel.

Toyota joins a number of other companies forming AI-focused venture capital funds including Baidu, which established theirs last year, and Google's Gradient Ventures, which was announced today. In a statement TRI CEO Gill Pratt said, "TRI is growing quickly, and we recognize the critical importance of expanding our collaboration with the world's brightest entrepreneurial talent. This venture is a major step towards our mission to empower talented entrepreneurs who share Toyota's commitment to enhancing the quality of human life."

View original post here:

Toyota's $100 million fund will back AI, robotics startups - Engadget

Over the last several years, investors looking to benefit from the ongoing developments in the field of AI needed to look no further than graphics processing pioneer NVIDIA Corporation (NASDAQ:NVDA). The massive parallel computing capability that made GPU's the best choice for rendering images turned out to be just as effective for the training artificial intelligence (AI) systems. NVIDIA positioned itself to leverage that advantage and began optimizing processors specifically for that purpose.

For a time, the GPU giant had the field to itself and financial results soared. In its most recent quarter, NVIDIA grew revenue to $1.937 billion, an increase of 48% over the prior-year quarter, while net income of $507 million jumped 144% year over year. The stock has tripled in the last year, and its valuation has jumped as well. NVIDIA now trades at an astonishing 49 times trailing earnings, with an only slightly less expensive forward multiple of 42. At these levels, any actual or perceived failure to execute could bring the stock crashing down.

The good news is that investors looking to capitalize on the growing trend of AI can invest in a pioneer in the field that offers solid growth without the potential downside risk -- Google, a division of Alphabet Inc. (NASDAQ:GOOGL) (NASDAQ:GOOG).

Alphabet is a way to invest in AI for more risk-averse investors. Image source: Pixabay.

Google has been at the forefront of AI, and early research in deep learning, a specific discipline of AI, has led to advances in image recognition, language processing, and voice recognition. Suggesting the name of a friend to "tag" in a photo and the ability to ask questions of the virtual assistant on your smartphone are examples of the developments resulting from early successes in AI.

Google developed TensorFlow, its open-source AI framework that developers use to more easily build their own AI systems. The company also created the tensor processing unit (TPU), a specialized chip that delivers optimized performance, while achieving significant improvements in energy efficiency. These were previously only used in the execution or "inference" phase of running AI systems that had previously been trained using GPU's. Google recently revealed that its second-generation TPU is now capable of both the inference and training phases of AI systems, putting it into direct competition with NVIDIA's GPU's. Google has not yet announced plans to market the chip but is currently using the processor internally.

TPU's were instrumental in the historic win over a human champion in the ancient game of Go, one many thought too complex for a machine to master. These tools and technological advantages now underlie Google cloud computing systems and provide a catalyst for future growth. Market research company Gartner estimates that the cloud infrastructure-as-a-service (IaaS) market will top $34 billion in 2017, and grow to $71 billion by 2020. That market is currently dominated by Amazon.com,followed by Microsoft Corporation, but Google is third and closing fast.

The use of cloud services is becoming particularly relevant to development in the field of AI. The ability to train these systems requires the intersection of big data and vast computing power, and many companies don't possess the financial resources to develop AI programs from scratch. The ability to piggyback off the systems offered by cloud providers has been key to advancing the research capability of smaller companies.

AI will continue to revolutionize business. Image source: Getty Images.

It is difficult to quantify the future revenue potential of AI, but certain anecdotal evidence can provide insight. In 2014, Google acquired AI start-up DeepMind in a deal estimated at $600 million.At the time, Google sought to eek further energy efficiency from its already miserly data centers and applied DeepMind's AI to the task. By regulating cooling systems, windows, and servers, and controlling 120 condition-based variables, the company was able to reduce the amount of energy used for cooling by 40%.This cut Google's total power consumption by 15%, saving the company hundreds of millions of dollars.

While investors wait for the potential financial windfall that could result from AI, they can take heart that Google's principle business still thrives. In its most recent quarter, Alphabet increased revenue to $24.75 billion, up 22% over the prior-year quarter. Net income growth was similarly impressive at $5.4 billion, an increase of 29% year over year.

Alphabet stock is up 33% over the last year, respectable by any measure, but nowhere near the blistering pace of NVIDIA's 200% rise. Still, as the old saying goes "what goes up must come down." Google's development of the TPU illustrates a stark reality for NVIDIA. Should any processor or solution become generally available that improves the performance of the GPU, NVIDIA's future growth could slow considerably, and the stock will adjust to reflect that reality. Let the buyer beware.

Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool's board of directors. Teresa Kersten is an employee of LinkedIn and is a member of The Motley Fool's board of directors. LinkedIn is owned by Microsoft. Danny Vena owns shares of Alphabet (A shares) and Amazon. Danny Vena has the following options: long January 2018 $640 calls on Alphabet (C shares) and short January 2018 $650 calls on Alphabet (C shares). The Motley Fool owns shares of and recommends Alphabet (A shares), Alphabet (C shares), Amazon, and Nvidia. The Motley Fool has a disclosure policy.

Go here to read the rest:

Ignore NVIDIA Corporation: Here's a Better AI Stock - Motley Fool

I am very sure that the Arms AI is cheating, and I am not the only person who thinks so.

Im not saying this because losing to a games AI is a little embarrassing. Actually, Arms AI is remarkably robust. At higher levels, it is always one step ahead of me at every moment, much like AIs in fighting games like Tekken 7. The Arms AIs reflexes are slick and, frustratingly, its always readied some perfect counter for even my most clever moves. Thats normalits a computer. Not normal is how it appears to break the games physics engine to pummel me over and over again.

I swear Im not crazy. Theres a whole conversation going on in the Arms community about its AI. Last month, a Redditer noticed that when I activate my special while the CPUs arms are being extended, the CPU somehow immediately enters a block without having to retract the arms again, something I and many others noticed too. Commenters debated whether Arms AI is just as precise as other fighting games or whether, by doing stuff humans cant do, its shady and unfair.

I spent an hour looking for potentially game-breaking behavior while playing against Arms AIs ranging between levels five and seven. And heres what I found:

I cant wrap my head around the way it can magically retract its extended arms to block me, or how an arm can appear half-extended to foil a grab.

I cant understand why its arms nearly always take priority in situations where it should be more ambiguous. Playing against a high-level Arms AI, it feels like the game reluctantly cedes to you in fist-to-fist situations only when you land perfect direct hits (and the AIs fists seem to be much luckier).

Also confusing is how its arms seem to block my attacks after theyve hit.

Nintendo declined to comment when asked whether Arms AI is doing the 2017 equivalent of GameSharking, as they did when Competes Maddy Myers asked whether Mario Kart 8 Deluxes AI cheats, too.

Unlike in Super Smash Bros., players can actually grind against Arms (cheating) AI and level up in accuracy and dexterity. Thats good. But playing against a broken AI can also make the game less fun. An AIs difficulty should rely on proper strategy, not hacking.

Link:

I Swear, Arms' AI Must Be Cheating - Kotaku

In Brief Researchers at DeepMind have published a paper illustrating how they are teaching artificially intelligent computer agents to traverse alien environments. While the results are slightly goofy, they represent a major step forward on the path to autonomous AI movement.

Googles artificial intelligence (AI) subsidiary DeepMind hasreleased a paperdetailing how itsAI agents have taught themselves to navigate complex virtual environments, and the results are weird, wonderful, and often extremely funny.

The agents in the simulations were programmed with a set of sensors these allowed them to know things like when they were upright or if their leg was bent and a drive to continue moving forward. Everything else that you see in the video the agents jumping, running, using knees to scale obstacles, etc. is the result of the AI working out how best to continue moving forward through reinforcement learning.

The complexity of the agents movements is a testament to how far AI has come in recent years. While agents in simulations like these often break down when faced with unfamiliar environments, DeepMinds haveutilized startlingly sophisticated movements to traverse obstacles.

These agile AIs arent the first to impress, though. A DeepMind AI has previously illustrated super-humanperformance levels on an object recognition task, anda team at the University of Cambridge has developed an AI system capable of performing more abstractly cerebral tasks, such asreading emotions and detecting pain levels.

The groundwork being laid by experiments such as these is pivotal to the integration of AI into society. Eventually, researchers will be able to incorporate these advancements into the programming of future AI robots, which will be able to navigate around your home or the streets, ushering in the age of truly seamless robot/human interaction.

More here:

DeepMind Has Taught an AI to Do Something Quite Remarkable - Futurism