Monthly Archives: August 2017

The ultimate idea is not to replace artists but to give them tools that allow them to create in entirely new ways.

MOUNTAIN VIEW, Calif. In the mid-1990s, Douglas Eck worked as a database programmer in Albuquerque, New Mexico, while moonlighting as a musician. After a day spent writing computer code inside a lab run by the Department of Energy, he would take the stage at a local juke joint, playing what he calls punk-influenced bluegrass Johnny Rotten crossed with Johnny Cash. But what he really wanted to do was combine his days and nights, and build machines that could make their own songs. My only goal in life was to mix AI and music, Eck said.

It was a naive ambition. Enrolling as a graduate student at Indiana University, in Bloomington, not far from where he grew up, he pitched the idea to Douglas Hofstadter, the cognitive scientist who wrote the Pulitzer Prize-winning book on minds and machines, Gdel, Escher, Bach: An Eternal Golden Braid.Hofstadter turned him down, adamant that even the latest artificial intelligence techniques were much too primitive.

But during the next two decades, working on the fringe of academia, Eck kept chasing the idea, and eventually, the AI caught up with his ambition.

Last spring, a few years after taking a research job at Google, Eck pitched the same idea he pitched Hofstadter all those years ago. The result is Project Magenta, a team of Google researchers who are teaching machines to create not only their own music but also to make so many other forms of art, including sketches, videos and jokes.

With its empire of smartphones, apps and internet services, Google is in the business of communication, and Eck sees Magenta as a natural extension of this work.

Its about creating new ways for people to communicate, he said during a recent interview inside the small two-story building here that serves as headquarters for Google AI research.

The project is part of a growing effort to generate art through a set of AI techniques that have only recently come of age. Called deep neural networks, these complex mathematical systems allow machines to learn specific behavior by analyzing vast amounts of data.

By looking for common patterns in millions of bicycle photos, for instance, a neural network can learn to recognize a bike. This is how Facebook identifies faces in online photos, how Android phones recognize commands spoken into phones, and how Microsoft Skype translates one language into another. But these complex systems can also create art. By analyzing a set of songs, for instance, they can learn to build similar sounds.

As Eck says, these systems are at least approaching the point still many, many years away when a machine can instantly build a new Beatles song or perhaps trillions of new Beatles songs, each sounding a lot like the music the Beatles themselves recorded, but also a little different.

But that end game as much a way of undermining art as creating it is not what he is after. There are so many other paths to explore beyond mere mimicry. The ultimate idea is not to replace artists but to give them tools that allow them to create in entirely new ways.

For centuries, orchestral conductors have layered sounds from various instruments atop one other. But this is different. Rather than layering sounds, Eck and his team are combining them to form something that did not exist before, creating new ways that artists can work.

Were making the next film camera, Eck said. Were making the next electric guitar.

Called NSynth, this particular project is only just getting off the ground. But across the worlds of both art and technology, many are already developing an appetite for building new art through neural networks and other AI techniques.

This work has exploded over the last few years, said Adam Ferris, a photographer and artist in Los Angeles. This is a totally new aesthetic.

In 2015, a separate team of researchers inside Google created DeepDream, a tool that uses neural networks to generate haunting, hallucinogenic imagescapes from existing photography, and this has spawned new art inside Google and out. If the tool analyzes a photo of a dog and finds a bit of fur that looks vaguely like an eyeball, it will enhance that bit of fur and then repeat the process. The result is a dog covered in swirling eyeballs.

At the same time, a number of artists like the well-known multimedia performance artist Trevor Paglen or the lesser-known Adam Ferris are exploring neural networks in other ways.

In January, Paglen gave a performance in an old maritime warehouse in San Francisco that explored the ethics of computer vision through neural networks that can track the way we look and move. While members of the avant-garde Kronos Quartet played onstage, for example, neural networks analyzed their expressions in real time, guessing at their emotions.

The tools are new, but the attitude is not. Allison Parrish, a New York University professor who builds software that generates poetry, points out that artists have been using computers to generate art since the 1950s. Much like as Jackson Pollock figured out a new way to paint by just opening the paint can and splashing it on the canvas beneath him, she said, these new computational techniques create a broader palette for artists.

A year ago, David Ha was a trader with Goldman Sachs in Tokyo. During his lunch breaks he started toying with neural networks and posting the results to a blog under a pseudonym. Among other things, he built a neural network that learned to write its own Kanji, the logographic Chinese characters that are not so much written as drawn.

Soon, Eck and other Googlers spotted the blog, and now Ha is a researcher with Google Magenta. Through a project called SketchRNN, he is building neural networks that can draw.

By analyzing thousands of digital sketches made by ordinary people, these neural networks can learn to make images of things like pigs, trucks, boats or yoga poses. They do not copy what people have drawn. They learn to draw on their own, to mathematically identify what a pig drawing looks like. Then, you ask them to, say, draw a pig with a cats head, or to visually subtract a foot from a horse or sketch a truck that looks like a dog or build a boat from a few random squiggly lines.

Next to NSynth or DeepDream, these may seem less like tools that artists will use to build new works. But if you play with them, you realize that they are themselves art, living works built by Ha. AI is not just creating new kinds of art; it is creating new kinds of artists.

Read more:

AI is creating new types of art, and new types of artists - Seattle Times

We knew that Andrew Ng had more than just a series of deep learning courses up his sleeve when he announced the first phase of his deeplearning.ai last week. Its clear now that the turn of Ngs three part act is a $150 million venture capital fund, first noted by PEHub,targeting AI investments.

Ng, who formerly founded Googles Brain Team and served as chief scientist at Baidu has long evangelized the benefits AI could bring to the world. During an earlier conversation, Ng told me that his personal goal is to helpbring about an AI-powered society. It would follow that education via his deep learning classes is one step of that and providing capital and other resources is another.

2017 has been a particularly active year for starting AI-focused venture capital funds. In the last few months we have seen Google roll out Gradient Ventures,Basis Set Ventures hall in $136 million, Element.AI raise $102 million, Microsoft Ventures start its own AI fund and Toyota corral $100 million for AI investment.

Its unclear at this point how Ngs AI Fund will differentiate from the pack.Many of these funds are putting time and resources into securing data sets, technical mentors and advanced simulation tools to support the unique needs of AI startups. Of courseNgs name recognition and network should help ensure solid deal flow and enable Ng to poach and train talent for startups in need of scarce deep learning engineers.

Ive sent a note to Andrew and we will update this post if and when we get more details.

See more here:

Andrew Ng is raising a $150M AI Fund - TechCrunch

Elon Musk speaks in front of employees during the delivery of the first Tesla vehicle Model 3 on July 28, 2017.

Simmering tensions between the United States and North Korea have many people concerned about the possibility of nuclear war, but Elon Musk says the North Korean government doesn't pose as much of a threat to humanity as the rise of artificial intelligence (AI).

The SpaceX and Tesla CEO tweeted on Aug. 11: "If you're not concerned about AI safety, you should be. Vastly more risk than North Korea." The tweet was accompanied by a photo that features a pensive woman and a tag line that reads, "In the end the machines will win."

Concerns about the possibility of nuclear missile strikes have escalated in recent weeks, particularly after President Donald Trump and North Korean leader Kim Jong-un threatened each other with shows of force. The North Korean government even issued a statement saying it is "examining" plans for a missile strike near the U.S. territory of Guam.

But, Musk thinks humanity's most pressing concern could be closer to home.

The billionaire entrepreneur has been outspoken about the dangers of AI, and the need to take action before it's too late. In July, he spoke at the National Governors Association summer meeting and urged lawmakers to regulate AI now before it poses a grave threat to humanity.And in 2014, Musk said artificial intelligence is humanity's "biggest existential threat."

If you're not concerned about AI safety, you should be. Vastly more risk than North Korea. pic.twitter.com/2z0tiid0lc

Original article on Live Science.

Original post:

Elon Musk: AI Poses Bigger Threat to Humanity Than North Korea - Live Science

The Pentagon is lagging behind the tech industry when it comes to tapping artificial intelligence (AI) for national security, according to US defense secretary James Mattis. On a recent tour that included visits to Amazon and Google, Mattis spoke about his desire to better harness the technology for military purposes, according to a report from Wired.

"It's got to be better integrated by the Department of Defense, because I see many of the greatest advances out here on the West Coast in private industry," Mattis told Wired.

The tech sector has tapped AI for everything from data management to hiring to photography in recent years. The Defense Innovation Unit Experimental (DIUx), an organization founded in 2015 to work within the DoD, aims to make it easier for small tech companies to work with the DoD and the military. The unit has invested $100 million into 45 contracts, Wired noted, including those with companies developing autonomous drones that could investigate buildings during military raids, and a headset and microphone that can be mounted on a tooth.

Mattis told Wired that he hopes to see DIUx continue to gain expertise from the tech industry. "There's no doubt in my mind DIUx will continue to exist; it will grow in its influence on the Department of Defense," he said.

SEE: Defending against cyberwar: How the cybersecurity elite are working to prevent a digital apocalypse

However, in June, China announced plans to become a world leader in AI by 2030, investing heavily in the technology for its government, military, and companies to stay at the cutting edge and surpass their rivals. The US does not have a similar public, overarching strategy, Wired said. Further, the White House's budget proposal includes cuts to the National Science Foundation, which has long supported AI research.

A July report from Harvard's Belfer Center for Science and International Affairs, conducted on behalf of the director of the US Intelligence Advanced Research Projects Activity (IARPA), determined that "advances in machine learning and Artificial Intelligence (AI) represent a turning point in the use of automation in warfare," but that "many of the most transformative applications of AI have not yet been addressed."

And most AI research advances are occurring in the private sector and academia, with private sector funding dwarfing that of the US government, the report found.

Current AI capabilities could have a significant impact on national security, the report noted: For example, existing machine learning technology could allow for more automation in labor-intensive activities such as satellite imagery analysis and cyber defense.

Future progress in AI has the potential to transform national security technology, "on a par with nuclear weapons, aircraft, computers, and biotech," the Harvard report stated.

"The DoD needs to pursue AI solutions to stay competitive with its Chinese and Russian counterparts," said Roman Yampolskiy, director of the Cyber Security Laboratory at the University of Louisville. "Unfortunately, for the humanity that means development of killer robots, unsupervised drones and other mechanisms of killing people in an automated process. As we know all computer systems have bugs or can be hacked. What happens when our killer robots get hijacked by the enemy is something I am very concerned about."

At the enterprise level, 62% of security experts said they believe that AI will be weaponized and used for cyberattacks within the next 12 months, according to a recent survey from Cylance.

SEE: Special report: How to implement AI and machine learning

Machine learning in particular has seen some very important advances in recent years, as evidenced by work from tech giants such as Google and Amazon, including voice recognition, search correlation, and personalisation, according to Engin Kirda, professor of computer science at Northeastern University. This technology is also increasingly used in computer security applications, in distinguishing normal behavior from attack-related behavior, and detecting breaches, Kirda said.

"Seeing these advances, I think the Department of Defense is realizing the potential of machine learning (and AI in general), and is considering to invest more resources into catching up with some of the advances in consumer software," Kirda said. "That is a very smart thing to do, because it is clear that AI has great application potential for some of the application scenarios that the Department of Defense is interested in (e.g., anti-terror scenarios)."

From an IT standpoint, the DoD is the largest and most complex enterprise in the world, with over 10,000 networks and 4 million desktop computers, and millions of mobile computing devices, according to Bob Gourley, co-founder of the cyber security consultancy Cognitio and former CTO of the Defense Intelligence Agency. All of this IT exists to do one thing: Help execute the missions of national security.

"All DoD missions will always be human guided, but new AI approaches are already enhancing decision-making in military missions," Gourley said. "Machine learning algorithms are improving the ability of commanders to understand the environment and helping leaders assess best options. This will only improve."

However, leaders still lack the ability to choose the right AI for the right task, Gourley said. For example, thousands of models exist for search and discovery, and it is suboptimal to hard code a single algorithm into a solution. "Why not enable decision-makers to decide which code to use for the problem at hand?" Gourley said. "This will improve decision making and battlefield results."

Image: iStockphoto/ratpack223

Continue reading here:

US Sec. Mattis pushes military AI, experts warn of hijacked 'killer robots' - TechRepublic