Machines Teaching Each Other Could Be the Biggest Exponential Trend in AI

“The easier it is to communicate, the faster change happens.” – James Burke, Science Historian

During an October 2015 press conference announcing the autopilot feature of the Tesla Model S, which allowed the car to drive semi-autonomously, Tesla CEO Elon Musk said each driver would become an “expert trainer” for every Model S. Each car could improve its own autonomous features by learning from its driver, but more significantly, when one Tesla learned from its own driver—that knowledge could then be shared with every other Tesla vehicle.

As Fred Lambert with Electrik reported shortly after, Model S owners noticed how quickly the car’s driverless features were improving. In one example, Teslas were taking incorrect early exits along highways, forcing their owners to manually steer the car along the correct route. After just a few weeks, owners noted the cars were no longer taking premature exits.

“I find it remarkable that it is improving this rapidly,” said one Tesla owner.

Intelligent systems, like those powered by the latest round of machine learning software, aren’t just getting smarter: they’re getting smarter faster. Understanding the rate at which these systems develop can be a particularly challenging part of navigating technological change.

Ray Kurzweil has written extensively on the gaps in human understanding between what he calls the “intuitive linear” view of technological change and the “exponential” rate of change now taking place. Almost two decades after writing the influential essay on what he calls “The Law of Accelerating Returns”—a theory of evolutionary change concerned with the speed at which systems improve over time—connected devices are now sharing knowledge between themselves, escalating the speed at which they improve.

[Learn more about thinking exponentially and the Law of Accelerating Returns.]

“I think that this is perhaps the biggest exponential trend in AI,” said Hod Lipson, professor of mechanical engineering and data science at Columbia University, in a recent interview.

“All of the exponential technology trends have different ‘exponents,’” Lipson added. “But this one is potentially the biggest.”

According to Lipson, what we might call “machine teaching”—when devices communicate gained knowledge to one another—is a radical step up in the speed at which these systems improve.

“Sometimes it is cooperative, for example when one machine learns from another like a hive mind. But sometimes it is adversarial, like in an arms race between two systems playing chess against each other,” he said.

Lipson believes this way of developing AI is a big deal, in part, because it can bypass the need for training data.

“Data is the fuel of machine learning, but even for machines, some data is hard to get—it may be risky, slow, rare, or expensive. In those cases, machines can share experiences or create synthetic experiences for each other to augment or replace data. It turns out that this is not a minor effect, it actually is self-amplifying, and therefore exponential.”

Lipson sees the recent breakthrough from Google’s DeepMind, a project called AlphaGo Zero, as a stunning example of an AI learning without training data. Many are familiar with AlphaGo, the machine learning AI which became the world’s best Go a player after studying a massive training data-set comprised of millions of human Go moves. AlphaGo Zero, however, was able to beat even that Go-playing AI, simply by learning the rules of the game and playing by itself—no training data necessary. Then, just to show off, it beat the world’s best chess playing software after starting from scratch and training for only eight hours.

Now imagine thousands or more AlphaGo Zeroes instantaneously sharing their gained knowledge.

This isn’t just games though. Already, we’re seeing how it will have a major impact on the speed at which businesses can improve the performance of their devices.

One example is GE’s new industrial digital twin technology—a software simulation of a machine that models what is happening with the equipment. Think of it as a machine with its own self-image—which it can also share with technicians.

A steam turbine with a digital twin, for instance, can measure steam temperatures, rotor speeds, cold starts, and other data to predict breakdowns and warn technicians to prevent expensive repairs. The digital twins make these predictions by studying their own performance, but they also rely on models every other steam turbine has developed.

As machines begin to learn from their environments in new and powerful ways, their development is accelerated by communicating what they learn with each other. The collective intelligence of every GE turbine, spread across the planet, can accelerate each individual machine’s predictive ability. Where it may take one driverless car significant time to learn to navigate a particular city—one hundred driverless cars navigating that same city together, all sharing what they learn—can improve their algorithms in far less time.

As other AI-powered devices begin to leverage this shared knowledge transfer, we could see an even faster pace of development. So if you think things are developing quickly today, remember we’re only just getting started.

Image Credit: igor kisselev / Shutterstock.com

from Singularity Hub http://ift.tt/2mYnMyT

Advertisements

This Week’s Awesome Stories From Around the Web (Through January 20)

AUGMENTED REALITY

AR Has Inherited All the Promise and Hype of VR
Nick Statt | The Verge
“The full spectrum of AR was present on the floor of the show’s many gadget expos and booth halls, making it easy to see the parallels today with the VR of five years ago. VR these days has largely retreated into the background, with commercial products from three big consumer brands that have set a high bar and brushed away many of the subpar competitors and hopeful but impractical prototypes. In VR’s place, however, AR has taken up the mantle of the exciting new category we once reserved only for science fiction.”

ROBOTICS

Harvard’s milliDelta Robot Is Tiny and Scary Fast
Evan Ackerman | IEEE Spectrum
“Harvard’s delta robot takes all of this cleverness and shrinks it down into a fearsome little package. The 15 mm x 15 mm x 20 mm robot weighs just 430 milligrams, but it has a payload capacity of 1.3 grams. It can move around its 7 cubic millimeter workspace with a precision of about 5 micrometers. What’s really impressive, though, is the speed: It can reach velocities of 0.45 m/s, and accelerations of 215 m/s2, meaning that it can follow repeating patterns at a frequency of up to 75 Hz. Just watch…”

BIOTECH

U.S. Doctors Plan to Treat Cancer Patients Using CRISPR
Emily Mullin | MIT Technology Review
“Doctors at the University of Pennsylvania say they will use CRISPR to modify human immune cells so that they become expert cancer killers, according to plans posted this week to a directory of ongoing clinical trials. The study will enroll up to 18 patients fighting three different types of cancer—multiple myeloma, sarcoma, and melanoma—in what could become the first medical use of CRISPR outside China, where similar studies have been under way.”

SPACE

The Rocket That Will Take Elon’s Car to Mars Is About to Test Its Engines
Daniel Oberhaus | Motherboard
“SpaceX is scheduled to complete the first static fire of its Falcon Heavy, which the company hopes will one day carry astronauts to Mars. A static fire involves testing the rocket’s engines at full thrust for a few seconds while the rocket is held in place on the launch mount… According to SpaceX, the Falcon Heavy’s 27 Merlin engines pump out thrust that’s roughly equivalent to 18 Boeing 747 planes with their engines at full throttle.”

TECH

What’s at Stake With Amazon’s New HQ? Ask Newark
Issie Lapowsky | Wired
“Amazon’s picks for its so-called HQ2 are, with a few exceptions, already thriving. Newark is one of those exceptions. At 7.9 percent, Newark’s unemployment rate sits at roughly double the average of the other cities on the list. It has the highest poverty rate, too, with nearly one-third of its population living under the poverty line…This dichotomy of Newark’s extreme need and the proposal’s extreme generosity provides a striking illustration of what cities stand to gain—or lose—from Amazon’s decision.”

Image Credit: SpaceX / Flickr

from Singularity Hub http://ift.tt/2mSURLv

Life Inside China’s Total Surveillance State

http://img.youtube.com/vi/OQ5LnY21Hgc/0.jpg

China has turned the northwestern region of Xinjiang into a vast experiment in domestic surveillance. WSJ investigated what life is like in a place where one’s every move can be monitored with cutting-edge technology.

from Ethical Technology http://ift.tt/2DSoTHD

Remote-controlled DNA nanorobots could lead to the first nanorobotic production factory

http://img.youtube.com/vi/mY5192g1gQg/0.jpg

German researchers created a 55-nm-by-55-nm DNA-based molecular platform with a 25-nm-long robotic arm that can be actuated with externally applied electrical fields, under computer control. (credit: Enzo Kopperger et al./Science)

By powering a self-assembling DNA nanorobotic arm with electric fields, German scientists have achieved precise nanoscale movement that is at least five orders of magnitude (hundreds of thousands times) faster than previously reported DNA-driven robotic systems, they suggest today (Jan. 19) in the journal Science.

DNA origami has emerged as a powerful tool to build precise structures. But now, “Kopperger et al. make an impressive stride in this direction by creating a dynamic DNA origami structure that they can directly control from the macroscale with easily tunable electric fields—similar to a remote-controlled robot,” notes Björn Högberg of Karolinska Institutet in a related Perspective in Science, (p. 279).

The nanorobotic arm resembles the gearshift lever of a car. Controlled by an electric field (comparable to the car driver), short, single-stranded DNA serves as “latches” (yellow) to momentarily grab and lock the 25-nanometer-long arm into predefined “gear” positions. (credit: Enzo Kopperger et al./Science)

The new biohybrid nanorobotic systems could even act as a molecular mechanical memory (a sort of nanoscale version of the Babbage Analytical Engine), he notes. “With the capability to form long filaments with multiple DNA robot arms, the systems could also serve as a platform for new inventions in digital memory, nanoscale cargo transfer, and 3D printing of molecules.”

“The robot-arm system may be scaled up and integrated into larger hybrid systems by a combination of lithographic and self-assembly techniques,” according to the researchers. “Electrically clocked synthesis of molecules with a large number of robot arms in parallel could then be the first step toward the realization of a genuine nanorobotic production factory.”


Taking a different approach to a nanofactory, this “Productive Nanosystems: from Molecules to Superproducts” film — a collaborative project of animator and engineer John Burch and pioneer nanotechnologist K. Eric Drexler in 2005 — demonstrated key steps in a hypothetical process that converts simple molecules into a billion-CPU laptop computer. More here.


Abstract of A self-assembled nanoscale robotic arm controlled by electric fields

The use of dynamic, self-assembled DNA nanostructures in the context of nanorobotics requires fast and reliable actuation mechanisms. We therefore created a 55-nanometer–by–55-nanometer DNA-based molecular platform with an integrated robotic arm of length 25 nanometers, which can be extended to more than 400 nanometers and actuated with externally applied electrical fields. Precise, computer-controlled switching of the arm between arbitrary positions on the platform can be achieved within milliseconds, as demonstrated with single-pair Förster resonance energy transfer experiments and fluorescence microscopy. The arm can be used for electrically driven transport of molecules or nanoparticles over tens of nanometers, which is useful for the control of photonic and plasmonic processes. Application of piconewton forces by the robot arm is demonstrated in force-induced DNA duplex melting experiments.

from KurzweilAI http://ift.tt/2EXnzT2

Remote-controlled DNA nanorobots could lead to the first nanorobotic production factory

http://ift.tt/2FZoTpV

German researchers created a 55-nm-by-55-nm DNA-based molecular platform with a 25-nm-long robotic arm that can be actuated with externally applied electrical fields, under computer control. (credit: Enzo Kopperger et al./Science)

By powering a self-assembling DNA nanorobotic arm with electric fields, German scientists have achieved precise nanoscale movement that is at least five orders of magnitude (hundreds of thousands times) faster than previously reported DNA-driven robotic systems, they suggest today (Jan. 19) in the journal Science.

DNA origami has emerged as a powerful tool to build precise structures. But now, “Kopperger et al. make an impressive stride in this direction by creating a dynamic DNA origami structure that they can directly control from the macroscale with easily tunable electric fields—similar to a remote-controlled robot,” notes Björn Högberg of Karolinska Institutet in a related Perspective in Science, (p. 279).

The nanorobotic arm resembles the gearshift lever of a car. Controlled by an electric field (comparable to the car driver), short, single-stranded DNA serves as “latches” (yellow) to momentarily grab and lock the 25-nanometer-long arm into predefined “gear” positions. (credit: Enzo Kopperger et al./Science)

The new biohybrid nanorobotic systems could even act as a molecular mechanical memory (a sort of nanoscale version of the Babbage Analytical Engine), he notes. “With the capability to form long filaments with multiple DNA robot arms, the systems could also serve as a platform for new inventions in digital memory, nanoscale cargo transfer, and 3D printing of molecules.”

“The robot-arm system may be scaled up and integrated into larger hybrid systems by a combination of lithographic and self-assembly techniques,” according to the researchers. “Electrically clocked synthesis of molecules with a large number of robot arms in parallel could then be the first step toward the realization of a genuine nanorobotic production factory.”


Taking a different approach to a nanofactory, this “Productive Nanosystems: from Molecules to Superproducts” film — a collaborative project of animator and engineer John Burch and pioneer nanotechnologist K. Eric Drexler in 2005 — demonstrated key steps in a hypothetical process that converts simple molecules into a billion-CPU laptop computer. More here.


Abstract of A self-assembled nanoscale robotic arm controlled by electric fields

The use of dynamic, self-assembled DNA nanostructures in the context of nanorobotics requires fast and reliable actuation mechanisms. We therefore created a 55-nanometer–by–55-nanometer DNA-based molecular platform with an integrated robotic arm of length 25 nanometers, which can be extended to more than 400 nanometers and actuated with externally applied electrical fields. Precise, computer-controlled switching of the arm between arbitrary positions on the platform can be achieved within milliseconds, as demonstrated with single-pair Förster resonance energy transfer experiments and fluorescence microscopy. The arm can be used for electrically driven transport of molecules or nanoparticles over tens of nanometers, which is useful for the control of photonic and plasmonic processes. Application of piconewton forces by the robot arm is demonstrated in force-induced DNA duplex melting experiments.

from KurzweilAI » News http://ift.tt/2EXnzT2

Want Faster Data and a Cleaner Planet? Start Mining Asteroids

http://ift.tt/2CZWXnN

Mining asteroids might seem like the stuff of science fiction, but there are companies and a few governments already working hard to make it real. This should not be surprising: compared with the breathtaking bridges that engineers build on Earth, asteroid-mining is a simple, small-scale operation requiring only modest technological advances. If anything is lacking, it is the imagination to see how plausible it has become. I am afraid only that it might not arrive soon enough to address the urgent resource challenges that the world is facing right now.

As an academic researcher, I work with several asteroidmining companies to address that urgency. I depend on their funding, so there are trade secrets I cannot share. However, I can reveal the core reasons why I am optimistic about the business case for asteroid-mining, and what it will mean for our future.

Many people are skeptical of asteroid-mining because they imagine that the goal is to bring platinum back for sale in Earth’s metals market. Reporters repeatedly cite an irresistible statistic that the platinum in an asteroid can be worth trillions of US dollars, but anyone with an understanding of economics realizes that bringing home a huge stash of precious metal would crash the market, reducing the valuation of the asteroid.

On the other hand, if the plan is to dole out platinum in small quantities to keep the valuation high (as it is done in the diamond industry), then how could asteroid companies compete with terrestrial mining companies that benefit from a mature, low-cost terrestrial supply chain and transportation network?

This is exactly why platinum is not the objective of asteroid-mining. Instead, the first product from asteroids will be something much less obviously precious: water.

To rocket scientists, water is the raw material for propellant. Launching water from Earth into space consumes a lot of propellant, which makes the whole concept self-defeating. Fortunately, water is abundant in space, where it is much easier to move around. Water can be readily extracted from clay minerals in a common class of small bodies known as carbonaceous asteroids. Once separated from the minerals, the water can then be split by electricity (a process called electrolysis) into hydrogen and oxygen to make rocket propellant—the key ingredients of rocket fuel.

Using rocket propellant produced in space will reduce the cost of doing everything else in space, initiating a virtuous cycle for the off-Earth supply chain and transportation network. Before that can happen, though, we must find the customers who can get the whole process started.

Who will buy rocket fuel made from asteroid water? One concept is to sell it to telecommunications companies for boosting satellites into orbit. A decade ago, most satellites were launched with a small upper-stage rocket attached. The rocket initially lofts the satellite into geostationary transfer orbit, a highly elliptical orbit having perigee (the low point) just a few hundred kilometers above the Earth’s surface, and apogee (the high point) about 36,000 kilometers higher. The spacecraft coasts to apogee, where the rocket fires and circularizes the orbit so that the satellite can begin selling data to customers. The cost of the disposable upper-stage rocket is very high, however.

Today, most satellite owners place a lightweight electric thruster on the spacecraft instead. Such thrusters are cheaper and more efficient, but very weak. It takes 6 to 12 months for satellites to reach final orbit. Time is money, so this delay still costs the satellite-owners hundreds of millions of dollars in lost revenues.

Asteroid-mining will provide a third option. A mining company will sell water to an in-space transportation company, which will use it to refuel a space tug parked in Earth orbit. The tug will dock with the newly launched satellite in geostationary transfer orbit, and boost it to the final orbit quickly, within a day.

According to our calculations, the total cost for this service, including capital recovery, finance charges, insurance and profit for all parties, will be less than the lost revenues of the current method, so that means there is a business case. The only concern is whether there are enough early customers to get the service established.

Here is where the national space agencies like NASA can help. If they develop an in-space refueling depot to lower their costs for exploring the Moon or Mars, and if they give out commercial contracts for some of this space water, they will lower the capital investment and risk for the new mining companies. In this way, government agencies can ensure the earlier success of private space industry. This is a legitimate role for government because taxpayers will greatly benefit.

An asteroid-mining infrastructure could help to solve a major impending resource problem. Within a decade or two, the current system of satellites and fiber optics will not be able to keep up with the demand for wireless and internet data. I know of no solution apart from building antennas in space that are too large to launch on rockets, because nothing else scales up quickly enough to meet the data needs that will grow exponentially through to the end of the century. Metal from asteroids will not be sold on Earth, where it would be too expensive. It will remain in space, transmitting precious data down into the digital market.

Similar arguments can be made that generating solar energy in space will, by sometime this century, be cheaper than generating energy on Earth through any known method. The energy might then be beamed to the ground via microwaves. Moving most of the energy sector into space will unburden the planet of the environmental impacts of energy generation, along with the entire supply chain that supports it. Even wind and solar disrupt large areas of land.

Off-planet energy generation could eliminate one-quarter of the human industrial footprint by 2100, by some estimates. This does not even take into account the exponentially growing energy footprint of computer manufacturing and operation, which is terrifying from an environmental perspective.

Note that none of these ideas involves bringing asteroid materials back for sale on Earth. The real value of space-based mining will be to create a space-based industry that benefits all of us. The primary import from space will be massless photons carrying data and energy.

The important point our government leaders should understand is that investing in space-mining is a safe bet on our future, one of the safest they can make. NASA and the other space agencies will get more science and exploration, plus greater geopolitical presence, for less cost than their current way of doing business. Saving the Earth and improving our quality of life might simply be side effects we get for free.Aeon counter – do not remove

This article was originally published at Aeon and has been republished under Creative Commons.

Image Credit: Dotted Yeti / Shutterstock.com

from Singularity Hub http://ift.tt/2Dtqs0D

A Blueprint for Building a Collaborative Startup Culture

Culture is driven from the top down. Yet most founders have a difficult time articulating their culture beyond the amenities their startup offers: Catered meals. Pet-friendly workspaces. Unlimited paid time off. Etc.

But a company’s culture is deeper than that. It’s an ideology, a way of being, and a mindset. It’s the intangibles that get your talent engaged, motivated, and wanting to do their best work.

So What Is Collaboration?

“If you want to go fast, go alone. If you want to go far, go together.” – African proverb

Bringing people together to arrive at a solution faster and easier than had it been done alone: That’s collaboration in a nutshell.

You innovate, solve problems together, bring together cross-functional expertise and knowledge to build an app, see a new solution, or tackle a coding issue. That’s how most startups see collaboration. But it’s also the ability to bring people together to talk about difficult topics. Conflict is inevitable, and in startups, where you’re bringing different people together to work on projects, personality clashes are all too common. So building that collaborative startup culture also requires seeing conflict as a positive thing, building out the values and processes necessary to leverage its beneficial qualities.

Conflict leads to better solutions, higher-quality work, and hyper-motivated employees, because disagreement is seen as a way of generating solutions that may not be readily available.

To create that culture of collaboration and push past resistance, startup founders should do the following.

1. Get Clear on Your Purpose and Values

Can you recite your company’s values from memory? Do you know why you’re tackling the problem you are, or why your solution is important?

If you can’t get beyond the surface-level responses to these questions, chances are your employees won’t be able to either. Sit with these questions, and if you can’t find answers that passionately drive you forward, then go at it again, because this is what will inspire everyone in the company and motivate them when the going gets tough.

2. Encourage Bottom-Up Feedback Early On

How should your employees communicate their ideas, disagreements, and need for change? Will you set up “office hours,” an open-door policy, or an anonymous quarterly survey to gather feedback? This is all data for you to use.

Your employees are the ones on the front lines, the ones embodying your startup’s values and communicating its mission to customers. Let them tell you what’s working and what isn’t.

This is especially true in meetings. Take some intentional time before, during, or after the meeting has concluded to solicit feedback. If you see the one man who stays in himself not offer any feedback for the fifth time in a row, go to him one-on-one and encourage him to talk about anything he sees that needs improvement. The result may surprise you.

Employees who feel heard ultimately work harder and are more efficient in the long run.

3. Set Up Processes and Technologies That Empower People

Slack is ubiquitous in startups. The communication tool is how many startups keep each other in the know. But it may not work for your company, or there might be other communication tools your employees prefer. Either way, ask. Find the tools for your company that empower people to work effectively and are tailored to their needs.

4. Reward Collaboration Openly

If you see someone contributing an awesome idea, acknowledge it. If you see someone going above and beyond his or her role, acknowledge it.

Collaboration sometimes doesn’t come easy, and most employees may not know what collaboration actually is beyond just working together toward a shared goal. You’ll need to make the implicit explicit by acknowledging those collaborative endeavors that employees naturally do when they’re enthralled with their work. Bring up that stellar employee’s work at the next company meeting. Start a monthly employee mention where someone did a fabulous thing. Make a happy hour event all about something awesome a team did.

Create the culture that acknowledges the good while staying abreast of all the things your high-paced startup is doing well.

5. Have Disagreements Respectfully

Can your employees openly disagree with you? Do they know how to bring up opposing opinions respectfully? You’ll know you can successfully collaborate with someone if you can have open disagreement.

That doesn’t mean fighting it out. It means learning how to hash things out in a collaborative way so that whatever solution is presented integrates everyone’s expertise and knowledge base. Disagreements, if carried out respectfully, put everyone’s cards on the table and present solutions that might not have been apparent before.

6. Harness Transparency and Build Trust

Startups are high-risk environments. When the company hits roadblocks, how does it communicate them to others? When funding doesn’t come through, do you tell your employees or not? When your current runway is about to expire, who is in on the information?

Startup founders can get into trouble when they don’t communicate certain facts to their employees. But it’s all a balancing act too. The amount of transparency founders exhibit will be the same amount their employees exhibit. In the high-risk environment of startups, that is the social currency needed to make your business a success.

7. Leverage Individual Employee Talents

CFO: “What happens if we invest in developing our employees, and they leave?” 

CEO: “What happens if we don’t, and they stay?”

As people start to work together, untapped talents become apparent. Leverage those talents by developing a method for spotting them. Give employees opportunities to vocalize them. For example, if you have a marketing director who’s consistently bringing in outside business, notice that. Develop her strategic partnerships skills and leverage them. A collaborative startup culture knows how to train employees for growth, retain them, and use their skills collaboratively toward an intended purpose.

Bottom Line: Creating a collaborative startup culture starts from the top, and is heavily influenced by the early decisions you make when your startup is scaling. Develop a culture that rewards innovation and disagreement and empowers its employees. That’s how a high-paced startup fuses collaboration into all areas of its culture to ultimately create an environment where everyone can thrive.

Image Credit: GaudiLab / Shutterstock.com

from Singularity Hub http://ift.tt/2mUuBBs