Tag Archives: under

#431171 SceneScan: Real-Time 3D Depth Sensing ...

Nerian Introduces a High-Performance Successor for the Proven SP1 System
Stereo vision, which is the three-dimensional perception of our environment with two sensors likeour eyes, is a well-known technology. As a passive method – there is no need to emit light in thevisible or invisible spectral range – this technology can open up new possibilities for three dimensional perception, even under difficult conditions.
But as often, the devil is in the details: for most applications, the software implementation withstandard PCs, but also with graphics processors, is too slow. Another complicating factor is thatthese hardware platforms are expensive and not energy-efficient. The solution is to instead usespecialized hardware for image processing. A programmable logic device – a so-called FPGA – cangreatly accelerate the image processing.
As a technology leader, Nerian Vision Technologies has been following this path successfully forthe past two years with the SP1 stereo vision system, which has enabled completely newapplications in the fields of robotics, automation technology, medical technology, autonomousdriving and other domains. Now the company introduces two successors:
SceneScan and SceneScan Pro. Real eye-catchers in a double sense: stereo vision in an elegant design!But more important is, of course, the significantly improved inner workings of the two new modelsin comparison to their predecessor. The new hardware allows processing rates of up to 100 framesper second at resolutions of up to 3 megapixels, which leaves the SP1 far behind:
Photo Credit: Nerian Vision Technologies – www.nerian.com

The table illustrates the difference: while SceneScan Pro has the highest possible computing powerand is designed for the most demanding applications, SceneScan has been cost-reduced forapplications with lower requirements. The customer can thus optimize his embedded vision solution both in terms of costs and technology.
The new duo is completed by Nerian’s proven Karmin stereo cameras. Of course, industrialUSB3Vision cameras by other manufacturers are also supported.This combination not only supports the above-mentioned applications even better, but alsofacilitates completely new and innovative ones. If required, customer-specific adaptations are alsopossible.
ContactNerian Vision TechnologiesOwner: Dr. Konstantin SchauweckerGotenstr. 970771 Leinfelden-EchterdingenGermanyPhone: +49 711 / 2195 9414Email: service@nerian.comWebsite: http://nerian.com
Press Release Authored By: Nerian Vision Technologies
Photo Credit: Nerian Vision Technologies – www.nerian.com
The post SceneScan: Real-Time 3D Depth Sensing Through Stereo Vision appeared first on Roboticmagazine. Continue reading

Posted in Human Robots | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , | Leave a comment

#431000 Japan’s SoftBank Is Investing Billions ...

Remember the 1980s movie Brewster’s Millions, in which a minor league baseball pitcher (played by Richard Pryor) must spend $30 million in 30 days to inherit $300 million? Pryor goes on an epic spending spree for a bigger payoff down the road.
One of the world’s biggest public companies is making that film look like a weekend in the Hamptons. Japan’s SoftBank Group, led by its indefatigable CEO Masayoshi Son, is shooting to invest $100 billion over the next five years toward what the company calls the information revolution.
The newly-created SoftBank Vision Fund, with a handful of key investors, appears ready to almost single-handedly hack the technology revolution. Announced only last year, the fund had its first major close in May with $93 billion in committed capital. The rest of the money is expected to be raised this year.
The fund is unprecedented. Data firm CB Insights notes that the SoftBank Vision Fund, if and when it hits the $100 billion mark, will equal the total amount that VC-backed companies received in all of 2016—$100.8 billion across 8,372 deals globally.
The money will go toward both billion-dollar corporations and startups, with a minimum $100 million buy-in. The focus is on core technologies like artificial intelligence, robotics and the Internet of Things.
Aside from being Japan’s richest man, Son is also a futurist who has predicted the singularity, the moment in time when machines will become smarter than humans and technology will progress exponentially. Son pegs the date as 2047. He appears to be hedging that bet in the biggest way possible.
Show Me the Money
Ostensibly a telecommunications company, SoftBank Group was founded in 1981 and started investing in internet technologies by the mid-1990s. Son infamously lost about $70 billion of his own fortune after the dot-com bubble burst around 2001. The company itself has a market cap of nearly $90 billion today, about half of where it was during the heydays of the internet boom.
The ups and downs did nothing to slake the company’s thirst for technology. It has made nine acquisitions and more than 130 investments since 1995. In 2017 alone, SoftBank has poured billions into nearly 30 companies and acquired three others. Some of those investments are being transferred to the massive SoftBank Vision Fund.
SoftBank is not going it alone with the new fund. More than half of the money—$60 billion—comes via the Middle East through Saudi Arabia’s Public Investment Fund ($45 billion) and Abu Dhabi’s Mubadala Investment Company ($15 billion). Other players at the table include Apple, Qualcomm, Sharp, Foxconn, and Oracle.
During a company conference in August, Son notes the SoftBank Vision Fund is not just about making money. “We don’t just want to be an investor just for the money game,” he says through a translator. “We want to make the information revolution. To do the information revolution, you can’t do it by yourself; you need a lot of synergy.”
Off to the Races
The fund has wasted little time creating that synergy. In July, its first official investment, not surprisingly, went to a company that specializes in artificial intelligence for robots—Brain Corp. The San Diego-based startup uses AI to turn manual machines into self-driving robots that navigate their environments autonomously. The first commercial application appears to be a really smart commercial-grade version that crosses a Roomba and Zamboni.

A second investment in July was a bit more surprising. SoftBank and its fund partners led a $200 million mega-round for Plenty, an agricultural tech company that promises to reshape farming by going vertical. Using IoT sensors and machine learning, Plenty claims its urban vertical farms can produce 350 times more vegetables than a conventional farm using 1 percent of the water.
Round Two
The spending spree continued into August.
The SoftBank Vision Fund led a $1.1 billion investment into a little-known biotechnology company called Roivant Sciences that goes dumpster diving for abandoned drugs and then creates subsidiaries around each therapy. For example, Axovant Sciences is devoted to neurology while Urovant focuses on urology. TechCrunch reports that Roivant is also creating a tech-focused subsidiary, called Datavant, that will use AI for drug discovery and other healthcare initiatives, such as designing clinical trials.
The AI angle may partly explain SoftBank’s interest in backing the biggest private placement in healthcare to date.
Also in August, SoftBank Vision Fund led a mix of $2.5 billion in primary and secondary capital investments into India’s largest private company in what was touted as the largest single investment in a private Indian company. Flipkart is an e-commerce company in the mold of Amazon.
The fund tacked on a $250 million investment round in August to Kabbage, an Atlanta-based startup in the alt-lending sector for small businesses. It ended big with a $4.4 billion investment into a co-working company called WeWork.
Betterment of Humanity
And those investments only include companies that SoftBank Vision Fund has backed directly.
SoftBank the company will offer—or has already turned over—previous investments to the Vision Fund in more than a half-dozen companies. Those assets include its shares in Nvidia, which produces chips for AI applications, and its first serious foray into autonomous driving with Nauto, a California startup that uses AI and high-tech cameras to retrofit vehicles to improve driving safety. The more miles the AI logs, the more it learns about safe and unsafe driving behaviors.
Other recent acquisitions, such as Boston Dynamics, a well-known US robotics company owned briefly by Google’s parent company Alphabet, will remain under the SoftBank Group umbrella for now.

This spending spree begs the question: What is the overall vision behind the SoftBank’s relentless pursuit of technology companies? A spokesperson for SoftBank told Singularity Hub that the “common thread among all of these companies is that they are creating the foundational platforms for the next stage of the information revolution.All of the companies, he adds, share SoftBank’s criteria of working toward “the betterment of humanity.”
While the SoftBank portfolio is diverse, from agtech to fintech to biotech, it’s obvious that SoftBank is betting on technologies that will connect the world in new and amazing ways. For instance, it wrote a $1 billion check last year in support of OneWeb, which aims to launch 900 satellites to bring internet to everyone on the planet. (It will also be turned over to the SoftBank Vision Fund.)
SoftBank also led a half-billion equity investment round earlier this year in a UK company called Improbable, which employs cloud-based distributed computing to create virtual worlds for gaming. The next step for the company is massive simulations of the real world that supports simultaneous users who can experience the same environment together(and another candidate for the SoftBank Vision Fund.)
Even something as seemingly low-tech as WeWork, which provides a desk or office in locations around the world, points toward a more connected planet.
In the end, the singularity is about bringing humanity together through technology. No one said it would be easy—or cheap.
Stock Media provided by xackerz / Pond5 Continue reading

Posted in Human Robots | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , | Comments Off on Japan’s SoftBank Is Investing Billions ...

#430855 Why Education Is the Hardest Sector of ...

We’ve all heard the warning cries: automation will disrupt entire industries and put millions of people out of jobs. In fact, up to 45 percent of existing jobs can be automated using current technology.
However, this may not necessarily apply to the education sector. After a detailed analysis of more than 2,000-plus work activities for more than 800 occupations, a report by McKinsey & Co states that of all the sectors examined, “…the technical feasibility of automation is lowest in education.”
There is no doubt that technological trends will have a powerful impact on global education, both by improving the overall learning experience and by increasing global access to education. Massive open online courses (MOOCs), chatbot tutors, and AI-powered lesson plans are just a few examples of the digital transformation in global education. But will robots and artificial intelligence ever fully replace teachers?
The Most Difficult Sector to Automate
While various tasks revolving around education—like administrative tasks or facilities maintenance—are open to automation, teaching itself is not.
Effective education involves more than just transfer of information from a teacher to a student. Good teaching requires complex social interactions and adaptation to the individual student’s learning needs. An effective teacher is not just responsive to each student’s strengths and weaknesses, but is also empathetic towards the student’s state of mind. It’s about maximizing human potential.
Furthermore, students don’t just rely on effective teachers to teach them the course material, but also as a source of life guidance and career mentorship. Deep and meaningful human interaction is crucial and is something that is very difficult, if not impossible, to automate.
Automating teaching is an example of a task that would require artificial general intelligence (as opposed to narrow or specific intelligence). In other words, this is the kind of task that would require an AI that understands natural human language, can be empathetic towards emotions, plan, strategize and make impactful decisions under unpredictable circumstances.
This would be the kind of machine that can do anything a human can do, and it doesn’t exist—at least, not yet.
We’re Getting There
Let’s not forget how quickly AI is evolving. Just because it’s difficult to fully automate teaching, it doesn’t mean the world’s leading AI experts aren’t trying.
Meet Jill Watson, the teaching assistant from Georgia Institute of Technology. Watson isn’t your average TA. She’s an IBM-powered artificial intelligence that is being implemented in universities around the world. Watson is able to answer students’ questions with 97 percent certainty.
Technologies like this also have applications in grading and providing feedback. Some AI algorithms are being trained and refined to perform automatic essay scoring. One project has achieved a 0.945 correlation with human graders.
All of this will have a remarkable impact on online education as we know it and dramatically increase online student retention rates.

Any student with a smartphone can access a wealth of information and free courses from universities around the world. MOOCs have allowed valuable courses to become available to millions of students. But at the moment, not all participants can receive customized feedback for their work. Currently, this is limited by manpower, but in the future that may not be the case.
What chatbots like Jill Watson allow is the opportunity for hundreds of thousands, if not millions, of students to have their work reviewed and all their questions answered at a minimal cost.
AI algorithms also have a significant role to play in personalization of education. Every student is unique and has a different set of strengths and weaknesses. Data analysis can be used to improve individual student results, assess each student’s strengths and weaknesses, and create mass-customized programs. Algorithms can analyze student data and consequently make flexible programs that adapt to the learner based on real-time feedback. According to the McKinsey Global Institute, all of this data in education could unlock between $900 billion and $1.2 trillion in global economic value.
Beyond Automated Teaching
It’s important to recognize that technological automation alone won’t fix the many issues in our global education system today. Dominated by outdated curricula, standardized tests, and an emphasis on short-term knowledge, many experts are calling for a transformation of how we teach.
It is not enough to simply automate the process. We can have a completely digital learning experience that continues to focus on outdated skills and fails to prepare students for the future. In other words, we must not only be innovative with our automation capabilities, but also with educational content, strategy, and policies.
Are we equipping students with the most important survival skills? Are we inspiring young minds to create a better future? Are we meeting the unique learning needs of each and every student? There’s no point automating and digitizing a system that is already flawed. We need to ensure the system that is being digitized is itself being transformed for the better.
Stock Media provided by davincidig / Pond5 Continue reading

Posted in Human Robots | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , | Comments Off on Why Education Is the Hardest Sector of ...

#430830 Biocomputers Made From Cells Can Now ...

When it comes to biomolecules, RNA doesn’t get a lot of love.
Maybe you haven’t even heard of the silent workhorse. RNA is the cell’s de facto translator: like a game of telephone, RNA takes DNA’s genetic code to a cellular factory called ribosomes. There, the cell makes proteins based on RNA’s message.
But RNA isn’t just a middleman. It controls what proteins are formed. Because proteins wiz around the cell completing all sorts of important processes, you can say that RNA is the gatekeeper: no RNA message, no proteins, no life.
In a new study published in Nature, RNA finally took center stage. By adding bits of genetic material to the E. Coli bacteria, a team of biohackers at the Wyss Institute hijacked the organism’s RNA messengers so that they only spring into action following certain inputs.
The result? A bacterial biocomputer capable of performing 12-input logic operations—AND, OR, and NOT—following specific inputs. Rather than outputting 0s and 1s, these biocircuits produce results based on the presence or absence of proteins and other molecules.
“It’s the greatest number of inputs in a circuit that a cell has been able to process,” says study author Dr. Alexander Green at Arizona State University. “To be able to analyze those signals and make a decision is the big advance here.”
When given a specific set of inputs, the bacteria spit out a protein that made them glow neon green under fluorescent light.
But synthetic biology promises far more than just a party trick—by tinkering with a cell’s RNA repertoire, scientists may one day coax them to photosynthesize, produce expensive drugs on the fly, or diagnose and hunt down rogue tumor cells.
Illustration of an RNA-based ‘ribocomputing’ device that makes logic-based decisions in living cells. The long gate RNA (blue) detects the binding of an input RNA (red). The ribosome (purple/mauve) reads the gate RNA to produce an output protein. Image Credit: Alexander Green / Arizona State University
The software of life
This isn’t the first time that scientists hijacked life’s algorithms to reprogram cells into nanocomputing systems. Previous work has already introduced to the world yeast cells that can make anti-malaria drugs from sugar or mammalian cells that can perform Boolean logic.
Yet circuits with multiple inputs and outputs remain hard to program. The reason is this: synthetic biologists have traditionally focused on snipping, fusing, or otherwise arranging a cell’s DNA to produce the outcomes they want.
But DNA is two steps removed from proteins, and tinkering with life’s code often leads to unexpected consequences. For one, the cell may not even accept and produce the extra bits of DNA code. For another, the added code, when transformed into proteins, may not act accordingly in the crowded and ever-changing environment of the cell.
What’s more, tinkering with one gene is often not enough to program an entirely new circuit. Scientists often need to amp up or shut down the activity of multiple genes, or multiple biological “modules” each made up of tens or hundreds of genes.
It’s like trying to fit new Lego pieces in a specific order into a room full of Lego constructions. Each new piece has the potential to wander off track and click onto something it’s not supposed to touch.
Getting every moving component to work in sync—as you might have guessed—is a giant headache.
The RNA way
With “ribocomputing,” Green and colleagues set off to tackle a main problem in synthetic biology: predictability.
Named after the “R (ribo)” in “RNA,” the method grew out of an idea that first struck Green back in 2012.
“The synthetic biological circuits to date have relied heavily on protein-based regulators that are difficult to scale up,” Green wrote at the time. We only have a limited handful of “designable parts” that work well, and these circuits require significant resources to encode and operate, he explains.
RNA, in comparison, is a lot more predictable. Like its more famous sibling DNA, RNA is composed of units that come in four different flavors: A, G, C, and U. Although RNA is only single-stranded, rather than the double helix for which DNA is known for, it can bind short DNA-like sequences in a very predictable manner: Gs always match up with Cs and As always with Us.
Because of this predictability, it’s possible to design RNA components that bind together perfectly. In other words, it reduces the chance that added RNA bits might go rogue in an unsuspecting cell.
Normally, once RNA is produced it immediately rushes to the ribosome—the cell’s protein-building factory. Think of it as a constantly “on” system.
However, Green and his team found a clever mechanism to slow them down. Dubbed the “toehold switch,” it works like this: the artificial RNA component is first incorporated into a chain of A, G, C, and U folded into a paperclip-like structure.
This blocks the RNA from accessing the ribosome. Because one RNA strand generally maps to one protein, the switch prevents that protein from ever getting made.
In this way, the switch is set to “off” by default—a “NOT” gate, in Boolean logic.
To activate the switch, the cell needs another component: a “trigger RNA,” which binds to the RNA toehold switch. This flips it on: the RNA grabs onto the ribosome, and bam—proteins.
BioLogic gates
String a few RNA switches together, with the activity of each one relying on the one before, and it forms an “AND” gate. Alternatively, if the activity of each switch is independent, that’s an “OR” gate.
“Basically, the toehold switches performed so well that we wanted to find a way to best exploit them for cellular applications,” says Green. They’re “kind of the equivalent of your first transistors,” he adds.
Once the team optimized the designs for different logic gates, they carefully condensed the switches into “gate RNA” molecules. These gate RNAs contain both codes for proteins and the logic operations needed to kickstart the process—a molecular logic circuit, so to speak.
If you’ve ever played around with an Arduino-controlled electrical circuit, you probably know the easiest way to test its function is with a light bulb.
That’s what the team did here, though with a biological bulb: green fluorescent protein, a light-sensing protein not normally present in bacteria that—when turned on—makes the microbugs glow neon green.
In a series of experiments, Green and his team genetically inserted gate RNAs into bacteria. Then, depending on the type of logical function, they added different combinations of trigger RNAs—the inputs.
When the input RNA matched up with its corresponding gate RNA, it flipped on the switch, causing the cell to light up.

Their most complex circuit contained five AND gates, five OR gates, and two NOTs—a 12-input ribocomputer that functioned exactly as designed.
That’s quite the achievement. “Everything is interacting with everything else and there are a million ways those interactions could flip the switch on accident,” says RNA researcher Dr. Julies Lucks at Northwestern University.
The specificity is thanks to RNA, the authors explain. Because RNAs bind to others so predictably, we can now design massive libraries of gate and trigger units to mix-and-match into all types of nano-biocomputers.
RNA BioNanobots
Although the technology doesn’t have any immediate applications, the team has high hopes.
For the first time, it’s now possible to massively scale up the process of programming new circuits into living cells. We’ve expanded the library of available biocomponents that can be used to reprogram life’s basic code, the authors say.
What’s more, when freeze-dried onto a piece of tissue paper, RNA keeps very well. We could potentially print RNA toehold switches onto paper that respond to viruses or to tumor cells, the authors say, essentially transforming the technology into highly accurate diagnostic platforms.
But Green’s hopes are even wilder for his RNA-based circuits.
“Because we’re using RNA, a universal molecule of life, we know these interactions can also work in other cells, so our method provides a general strategy that could be ported to other organisms,” he says.
Ultimately, the hope is to program neural network-like capabilities into the body’s other cells.
Imagine cells endowed with circuits capable of performing the kinds of computation the brain does, the authors say.
Perhaps one day, synthetic biology will transform our own cells into fully programmable entities, turning us all into biological cyborgs from the inside. How wild would that be?
Image Credit: Wyss Institute at Harvard University Continue reading

Posted in Human Robots | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , | Comments Off on Biocomputers Made From Cells Can Now ...

#430743 Teaching Machines to Understand, and ...

We humans are swamped with text. It’s not just news and other timely information: Regular people are drowning in legal documents. The problem is so bad we mostly ignore it. Every time a person uses a store’s loyalty rewards card or connects to an online service, his or her activities are governed by the equivalent of hundreds of pages of legalese. Most people pay no attention to these massive documents, often labeled “terms of service,” “user agreement,” or “privacy policy.”
These are just part of a much wider societal problem of information overload. There is so much data stored—exabytes of it, as much stored as has ever been spoken by people in all of human history—that it’s humanly impossible to read and interpret everything. Often, we narrow down our pool of information by choosing particular topics or issues to pay attention to. But it’s important to actually know the meaning and contents of the legal documents that govern how our data is stored and who can see it.
As computer science researchers, we are working on ways artificial intelligence algorithms could digest these massive texts and extract their meaning, presenting it in terms regular people can understand.
Can computers understand text?
Computers store data as 0s and 1s—data that cannot be directly understood by humans. They interpret these data as instructions for displaying text, sound, images, or videos that are meaningful to people. But can computers actually understand the language, not only presenting the words but also their meaning?
One way to find out is to ask computers to summarize their knowledge in ways that people can understand and find useful. It would be best if AI systems could process text quickly enough to help people make decisions as they are needed—for example, when you’re signing up for a new online service and are asked to agree with the site’s privacy policy.
What if a computerized assistant could digest all that legal jargon in a few seconds and highlight key points? Perhaps a user could even tell the automated assistant to pay particular attention to certain issues, like when an email address is shared, or whether search engines can index personal posts. Companies could use this capability, too, to analyze contracts or other lengthy documents.
To do this sort of work, we need to combine a range of AI technologies, including machine learning algorithms that take in large amounts of data and independently identify connections among them; knowledge representation techniques to express and interpret facts and rules about the world; speech recognition systems to convert spoken language to text; and human language comprehension programs that process the text and its context to determine what the user is telling the system to do.
Examining privacy policies
A modern internet-enabled life today more or less requires trusting for-profit companies with private information (like physical and email addresses, credit card numbers and bank account details) and personal data (photos and videos, email messages and location information).
These companies’ cloud-based systems typically keep multiple copies of users’ data as part of backup plans to prevent service outages. That means there are more potential targets—each data center must be securely protected both physically and electronically. Of course, internet companies recognize customers’ concerns and employ security teams to protect users’ data. But the specific and detailed legal obligations they undertake to do that are found in their impenetrable privacy policies. No regular human—and perhaps even no single attorney—can truly understand them.
In our study, we ask computers to summarize the terms and conditions regular users say they agree to when they click “Accept” or “Agree” buttons for online services. We downloaded the publicly available privacy policies of various internet companies, including Amazon AWS, Facebook, Google, HP, Oracle, PayPal, Salesforce, Snapchat, Twitter, and WhatsApp.
Summarizing meaning
Our software examines the text and uses information extraction techniques to identify key information specifying the legal rights, obligations and prohibitions identified in the document. It also uses linguistic analysis to identify whether each rule applies to the service provider, the user or a third-party entity, such as advertisers and marketing companies. Then it presents that information in clear, direct, human-readable statements.
For example, our system identified one aspect of Amazon’s privacy policy as telling a user, “You can choose not to provide certain information, but then you might not be able to take advantage of many of our features.” Another aspect of that policy was described as “We may also collect technical information to help us identify your device for fraud prevention and diagnostic purposes.”

We also found, with the help of the summarizing system, that privacy policies often include rules for third parties—companies that aren’t the service provider or the user—that people might not even know are involved in data storage and retrieval.
The largest number of rules in privacy policies—43 percent—apply to the company providing the service. Just under a quarter of the rules—24 percent—create obligations for users and customers. The rest of the rules govern behavior by third-party services or corporate partners, or could not be categorized by our system.

The next time you click the “I Agree” button, be aware that you may be agreeing to share your data with other hidden companies who will be analyzing it.
We are continuing to improve our ability to succinctly and accurately summarize complex privacy policy documents in ways that people can understand and use to access the risks associated with using a service.

This article was originally published on The Conversation. Read the original article. Continue reading

Posted in Human Robots | Tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , | Comments Off on Teaching Machines to Understand, and ...