Tag Archives: art

#434559 Can AI Tell the Difference Between a ...

Scarcely a day goes by without another headline about neural networks: some new task that deep learning algorithms can excel at, approaching or even surpassing human competence. As the application of this approach to computer vision has continued to improve, with algorithms capable of specialized recognition tasks like those found in medicine, the software is getting closer to widespread commercial use—for example, in self-driving cars. Our ability to recognize patterns is a huge part of human intelligence: if this can be done faster by machines, the consequences will be profound.

Yet, as ever with algorithms, there are deep concerns about their reliability, especially when we don’t know precisely how they work. State-of-the-art neural networks will confidently—and incorrectly—classify images that look like television static or abstract art as real-world objects like school-buses or armadillos. Specific algorithms could be targeted by “adversarial examples,” where adding an imperceptible amount of noise to an image can cause an algorithm to completely mistake one object for another. Machine learning experts enjoy constructing these images to trick advanced software, but if a self-driving car could be fooled by a few stickers, it might not be so fun for the passengers.

These difficulties are hard to smooth out in large part because we don’t have a great intuition for how these neural networks “see” and “recognize” objects. The main insight analyzing a trained network itself can give us is a series of statistical weights, associating certain groups of points with certain objects: this can be very difficult to interpret.

Now, new research from UCLA, published in the journal PLOS Computational Biology, is testing neural networks to understand the limits of their vision and the differences between computer vision and human vision. Nicholas Baker, Hongjing Lu, and Philip J. Kellman of UCLA, alongside Gennady Erlikhman of the University of Nevada, tested a deep convolutional neural network called VGG-19. This is state-of-the-art technology that is already outperforming humans on standardized tests like the ImageNet Large Scale Visual Recognition Challenge.

They found that, while humans tend to classify objects based on their overall (global) shape, deep neural networks are far more sensitive to the textures of objects, including local color gradients and the distribution of points on the object. This result helps explain why neural networks in image recognition make mistakes that no human ever would—and could allow for better designs in the future.

In the first experiment, a neural network was trained to sort images into 1 of 1,000 different categories. It was then presented with silhouettes of these images: all of the local information was lost, while only the outline of the object remained. Ordinarily, the trained neural net was capable of recognizing these objects, assigning more than 90% probability to the correct classification. Studying silhouettes, this dropped to 10%. While human observers could nearly always produce correct shape labels, the neural networks appeared almost insensitive to the overall shape of the images. On average, the correct object was ranked as the 209th most likely solution by the neural network, even though the overall shapes were an exact match.

A particularly striking example arose when they tried to get the neural networks to classify glass figurines of objects they could already recognize. While you or I might find it easy to identify a glass model of an otter or a polar bear, the neural network classified them as “oxygen mask” and “can opener” respectively. By presenting glass figurines, where the texture information that neural networks relied on for classifying objects is lost, the neural network was unable to recognize the objects by shape alone. The neural network was similarly hopeless at classifying objects based on drawings of their outline.

If you got one of these right, you’re better than state-of-the-art image recognition software. Image Credit: Nicholas Baker, Hongjing Lu, Gennady Erlikhman, Philip J. Kelman. “Deep convolutional networks do not classify based on global object shape.” Plos Computational Biology. 12/7/18. / CC BY 4.0
When the neural network was explicitly trained to recognize object silhouettes—given no information in the training data aside from the object outlines—the researchers found that slight distortions or “ripples” to the contour of the image were again enough to fool the AI, while humans paid them no mind.

The fact that neural networks seem to be insensitive to the overall shape of an object—relying instead on statistical similarities between local distributions of points—suggests a further experiment. What if you scrambled the images so that the overall shape was lost but local features were preserved? It turns out that the neural networks are far better and faster at recognizing scrambled versions of objects than outlines, even when humans struggle. Students could classify only 37% of the scrambled objects, while the neural network succeeded 83% of the time.

Humans vastly outperform machines at classifying object (a) as a bear, while the machine learning algorithm has few problems classifying the bear in figure (b). Image Credit: Nicholas Baker, Hongjing Lu, Gennady Erlikhman, Philip J. Kelman. “Deep convolutional networks do not classify based on global object shape.” Plos Computational Biology. 12/7/18. / CC BY 4.0
“This study shows these systems get the right answer in the images they were trained on without considering shape,” Kellman said. “For humans, overall shape is primary for object recognition, and identifying images by overall shape doesn’t seem to be in these deep learning systems at all.”

Naively, one might expect that—as the many layers of a neural network are modeled on connections between neurons in the brain and resemble the visual cortex specifically—the way computer vision operates must necessarily be similar to human vision. But this kind of research shows that, while the fundamental architecture might resemble that of the human brain, the resulting “mind” operates very differently.

Researchers can, increasingly, observe how the “neurons” in neural networks light up when exposed to stimuli and compare it to how biological systems respond to the same stimuli. Perhaps someday it might be possible to use these comparisons to understand how neural networks are “thinking” and how those responses differ from humans.

But, as yet, it takes a more experimental psychology to probe how neural networks and artificial intelligence algorithms perceive the world. The tests employed against the neural network are closer to how scientists might try to understand the senses of an animal or the developing brain of a young child rather than a piece of software.

By combining this experimental psychology with new neural network designs or error-correction techniques, it may be possible to make them even more reliable. Yet this research illustrates just how much we still don’t understand about the algorithms we’re creating and using: how they tick, how they make decisions, and how they’re different from us. As they play an ever-greater role in society, understanding the psychology of neural networks will be crucial if we want to use them wisely and effectively—and not end up missing the woods for the trees.

Image Credit: Irvan Pratama / Shutterstock.com Continue reading

Posted in Human Robots

#434260 The Most Surprising Tech Breakthroughs ...

Development across the entire information technology landscape certainly didn’t slow down this year. From CRISPR babies, to the rapid decline of the crypto markets, to a new robot on Mars, and discovery of subatomic particles that could change modern physics as we know it, there was no shortage of headline-grabbing breakthroughs and discoveries.

As 2018 comes to a close, we can pause and reflect on some of the biggest technology breakthroughs and scientific discoveries that occurred this year.

I reached out to a few Singularity University speakers and faculty across the various technology domains we cover asking what they thought the biggest breakthrough was in their area of expertise. The question posed was:

“What, in your opinion, was the biggest development in your area of focus this year? Or, what was the breakthrough you were most surprised by in 2018?”

I can share that for me, hands down, the most surprising development I came across in 2018 was learning that a publicly-traded company that was briefly valued at over $1 billion, and has over 12,000 employees and contractors spread around the world, has no physical office space and the entire business is run and operated from inside an online virtual world. This is Ready Player One stuff happening now.

For the rest, here’s what our experts had to say.

DIGITAL BIOLOGY
Dr. Tiffany Vora | Faculty Director and Vice Chair, Digital Biology and Medicine, Singularity University

“That’s easy: CRISPR babies. I knew it was technically possible, and I’ve spent two years predicting it would happen first in China. I knew it was just a matter of time but I failed to predict the lack of oversight, the dubious consent process, the paucity of publicly-available data, and the targeting of a disease that we already know how to prevent and treat and that the children were at low risk of anyway.

I’m not convinced that this counts as a technical breakthrough, since one of the girls probably isn’t immune to HIV, but it sure was a surprise.”

For more, read Dr. Vora’s summary of this recent stunning news from China regarding CRISPR-editing human embryos.

QUANTUM COMPUTING
Andrew Fursman | Co-Founder/CEO 1Qbit, Faculty, Quantum Computing, Singularity University

“There were two last-minute holiday season surprise quantum computing funding and technology breakthroughs:

First, right before the government shutdown, one priority legislative accomplishment will provide $1.2 billion in quantum computing research over the next five years. Second, there’s the rise of ions as a truly viable, scalable quantum computing architecture.”

*Read this Gizmodo profile on an exciting startup in the space to learn more about this type of quantum computing

ENERGY
Ramez Naam | Chair, Energy and Environmental Systems, Singularity University

“2018 had plenty of energy surprises. In solar, we saw unsubsidized prices in the sunny parts of the world at just over two cents per kwh, or less than half the price of new coal or gas electricity. In the US southwest and Texas, new solar is also now cheaper than new coal or gas. But even more shockingly, in Germany, which is one of the least sunny countries on earth (it gets less sunlight than Canada) the average bid for new solar in a 2018 auction was less than 5 US cents per kwh. That’s as cheap as new natural gas in the US, and far cheaper than coal, gas, or any other new electricity source in most of Europe.

In fact, it’s now cheaper in some parts of the world to build new solar or wind than to run existing coal plants. Think tank Carbon Tracker calculates that, over the next 10 years, it will become cheaper to build new wind or solar than to operate coal power in most of the world, including specifically the US, most of Europe, and—most importantly—India and the world’s dominant burner of coal, China.

Here comes the sun.”

GLOBAL GRAND CHALLENGES
Darlene Damm | Vice Chair, Faculty, Global Grand Challenges, Singularity University

“In 2018 we saw a lot of areas in the Global Grand Challenges move forward—advancements in robotic farming technology and cultured meat, low-cost 3D printed housing, more sophisticated types of online education expanding to every corner of the world, and governments creating new policies to deal with the ethics of the digital world. These were the areas we were watching and had predicted there would be change.

What most surprised me was to see young people, especially teenagers, start to harness technology in powerful ways and use it as a platform to make their voices heard and drive meaningful change in the world. In 2018 we saw teenagers speak out on a number of issues related to their well-being and launch digital movements around issues such as gun and school safety, global warming and environmental issues. We often talk about the harm technology can cause to young people, but on the flip side, it can be a very powerful tool for youth to start changing the world today and something I hope we see more of in the future.”

BUSINESS STRATEGY
Pascal Finette | Chair, Entrepreneurship and Open Innovation, Singularity University

“Without a doubt the rapid and massive adoption of AI, specifically deep learning, across industries, sectors, and organizations. What was a curiosity for most companies at the beginning of the year has quickly made its way into the boardroom and leadership meetings, and all the way down into the innovation and IT department’s agenda. You are hard-pressed to find a mid- to large-sized company today that is not experimenting or implementing AI in various aspects of its business.

On the slightly snarkier side of answering this question: The very rapid decline in interest in blockchain (and cryptocurrencies). The blockchain party was short, ferocious, and ended earlier than most would have anticipated, with a huge hangover for some. The good news—with the hot air dissipated, we can now focus on exploring the unique use cases where blockchain does indeed offer real advantages over centralized approaches.”

*Author note: snark is welcome and appreciated

ROBOTICS
Hod Lipson | Director, Creative Machines Lab, Columbia University

“The biggest surprise for me this year in robotics was learning dexterity. For decades, roboticists have been trying to understand and imitate dexterous manipulation. We humans seem to be able to manipulate objects with our fingers with incredible ease—imagine sifting through a bunch of keys in the dark, or tossing and catching a cube. And while there has been much progress in machine perception, dexterous manipulation remained elusive.

There seemed to be something almost magical in how we humans can physically manipulate the physical world around us. Decades of research in grasping and manipulation, and millions of dollars spent on robot-hand hardware development, has brought us little progress. But in late 2018, the Berkley OpenAI group demonstrated that this hurdle may finally succumb to machine learning as well. Given 200 years worth of practice, machines learned to manipulate a physical object with amazing fluidity. This might be the beginning of a new age for dexterous robotics.”

MACHINE LEARNING
Jeremy Howard | Founding Researcher, fast.ai, Founder/CEO, Enlitic, Faculty Data Science, Singularity University

“The biggest development in machine learning this year has been the development of effective natural language processing (NLP).

The New York Times published an article last month titled “Finally, a Machine That Can Finish Your Sentence,” which argued that NLP neural networks have reached a significant milestone in capability and speed of development. The “finishing your sentence” capability mentioned in the title refers to a type of neural network called a “language model,” which is literally a model that learns how to finish your sentences.

Earlier this year, two systems (one, called ELMO, is from the Allen Institute for AI, and the other, called ULMFiT, was developed by me and Sebastian Ruder) showed that such a model could be fine-tuned to dramatically improve the state-of-the-art in nearly every NLP task that researchers study. This work was further developed by OpenAI, which in turn was greatly scaled up by Google Brain, who created a system called BERT which reached human-level performance on some of NLP’s toughest challenges.

Over the next year, expect to see fine-tuned language models used for everything from understanding medical texts to building disruptive social media troll armies.”

DIGITAL MANUFACTURING
Andre Wegner | Founder/CEO Authentise, Chair, Digital Manufacturing, Singularity University

“Most surprising to me was the extent and speed at which the industry finally opened up.

While previously, only few 3D printing suppliers had APIs and knew what to do with them, 2018 saw nearly every OEM (or original equipment manufacturer) enabling data access and, even more surprisingly, shying away from proprietary standards and adopting MTConnect, as stalwarts such as 3D Systems and Stratasys have been. This means that in two to three years, data access to machines will be easy, commonplace, and free. The value will be in what is being done with that data.

Another example of this openness are the seemingly endless announcements of integrated workflows: GE’s announcement with most major software players to enable integrated solutions, EOS’s announcement with Siemens, and many more. It’s clear that all actors in the additive ecosystem have taken a step forward in terms of openness. The result is a faster pace of innovation, particularly in the software and data domains that are crucial to enabling comprehensive digital workflow to drive agile and resilient manufacturing.

I’m more optimistic we’ll achieve that now than I was at the end of 2017.”

SCIENCE AND DISCOVERY
Paul Saffo | Chair, Future Studies, Singularity University, Distinguished Visiting Scholar, Stanford Media-X Research Network

“The most important development in technology this year isn’t a technology, but rather the astonishing science surprises made possible by recent technology innovations. My short list includes the discovery of the “neptmoon”, a Neptune-scale moon circling a Jupiter-scale planet 8,000 lightyears from us; the successful deployment of the Mars InSight Lander a month ago; and the tantalizing ANITA detection (what could be a new subatomic particle which would in turn blow the standard model wide open). The highest use of invention is to support science discovery, because those discoveries in turn lead us to the future innovations that will improve the state of the world—and fire up our imaginations.”

ROBOTICS
Pablos Holman | Inventor, Hacker, Faculty, Singularity University

“Just five or ten years ago, if you’d asked any of us technologists “What is harder for robots? Eyes, or fingers?” We’d have all said eyes. Robots have extraordinary eyes now, but even in a surgical robot, the fingers are numb and don’t feel anything. Stanford robotics researchers have invented fingertips that can feel, and this will be a kingpin that allows robots to go everywhere they haven’t been yet.”

BLOCKCHAIN
Nathana Sharma | Blockchain, Policy, Law, and Ethics, Faculty, Singularity University

“2017 was the year of peak blockchain hype. 2018 has been a year of resetting expectations and technological development, even as the broader cryptocurrency markets have faced a winter. It’s now about seeing adoption and applications that people want and need to use rise. An incredible piece of news from December 2018 is that Facebook is developing a cryptocurrency for users to make payments through Whatsapp. That’s surprisingly fast mainstream adoption of this new technology, and indicates how powerful it is.”

ARTIFICIAL INTELLIGENCE
Neil Jacobstein | Chair, Artificial Intelligence and Robotics, Singularity University

“I think one of the most visible improvements in AI was illustrated by the Boston Dynamics Parkour video. This was not due to an improvement in brushless motors, accelerometers, or gears. It was due to improvements in AI algorithms and training data. To be fair, the video released was cherry-picked from numerous attempts, many of which ended with a crash. However, the fact that it could be accomplished at all in 2018 was a real win for both AI and robotics.”

NEUROSCIENCE
Divya Chander | Chair, Neuroscience, Singularity University

“2018 ushered in a new era of exponential trends in non-invasive brain modulation. Changing behavior or restoring function takes on a new meaning when invasive interfaces are no longer needed to manipulate neural circuitry. The end of 2018 saw two amazing announcements: the ability to grow neural organoids (mini-brains) in a dish from neural stem cells that started expressing electrical activity, mimicking the brain function of premature babies, and the first (known) application of CRISPR to genetically alter two fetuses grown through IVF. Although this was ostensibly to provide genetic resilience against HIV infections, imagine what would happen if we started tinkering with neural circuitry and intelligence.”

Image Credit: Yurchanka Siarhei / Shutterstock.com Continue reading

Posted in Human Robots

#433803 This Week’s Awesome Stories From ...

ARTIFICIAL INTELLIGENCE
The AI Cold War That Could Doom Us All
Nicholas Thompson | Wired
“At the dawn of a new stage in the digital revolution, the world’s two most powerful nations are rapidly retreating into positions of competitive isolation, like players across a Go board. …Is the arc of the digital revolution bending toward tyranny, and is there any way to stop it?”

LONGEVITY
Finally, the Drug That Keeps You Young
Stephen S. Hall | MIT Technology Review
“The other thing that has changed is that the field of senescence—and the recognition that senescent cells can be such drivers of aging—has finally gained acceptance. Whether those drugs will work in people is still an open question. But the first human trials are under way right now.”

SYNTHETIC BIOLOGY
Ginkgo Bioworks Is Turning Human Cells Into On-Demand Factories
Megan Molteni | Wired
“The biotech unicorn is already cranking out an impressive number of microbial biofactories that grow and multiply and burp out fragrances, fertilizers, and soon, psychoactive substances. And they do it at a fraction of the cost of traditional systems. But Kelly is thinking even bigger.”

CYBERNETICS
Thousands of Swedes Are Inserting Microchips Under Their Skin
Maddy Savage | NPR
“Around the size of a grain of rice, the chips typically are inserted into the skin just above each user’s thumb, using a syringe similar to that used for giving vaccinations. The procedure costs about $180. So many Swedes are lining up to get the microchips that the country’s main chipping company says it can’t keep up with the number of requests.”

ART
AI Art at Christie’s Sells for $432,500
Gabe Cohn | The New York Times
“Last Friday, a portrait produced by artificial intelligence was hanging at Christie’s New York opposite an Andy Warhol print and beside a bronze work by Roy Lichtenstein. On Thursday, it sold for well over double the price realized by both those pieces combined.”

ETHICS
Should a Self-Driving Car Kill the Baby or the Grandma? Depends on Where You’re From
Karen Hao | MIT Technology Review
“The researchers never predicted the experiment’s viral reception. Four years after the platform went live, millions of people in 233 countries and territories have logged 40 million decisions, making it one of the largest studies ever done on global moral preferences.”

TECHNOLOGY
The Rodney Brooks Rules for Predicting a Technology’s Success
Rodney Brooks | IEEE Spectrum
“Building electric cars and reusable rockets is fairly easy. Building a nuclear fusion reactor, flying cars, self-driving cars, or a Hyperloop system is very hard. What makes the difference?”

Image Source: spainter_vfx / Shutterstock.com Continue reading

Posted in Human Robots

#433799 The First Novel Written by AI Is ...

Last year, a novelist went on a road trip across the USA. The trip was an attempt to emulate Jack Kerouac—to go out on the road and find something essential to write about in the experience. There is, however, a key difference between this writer and anyone else talking your ear off in the bar. This writer is just a microphone, a GPS, and a camera hooked up to a laptop and a whole bunch of linear algebra.

People who are optimistic that artificial intelligence and machine learning won’t put us all out of a job say that human ingenuity and creativity will be difficult to imitate. The classic argument is that, just as machines freed us from repetitive manual tasks, machine learning will free us from repetitive intellectual tasks.

This leaves us free to spend more time on the rewarding aspects of our work, pursuing creative hobbies, spending time with loved ones, and generally being human.

In this worldview, creative works like a great novel or symphony, and the emotions they evoke, cannot be reduced to lines of code. Humans retain a dimension of superiority over algorithms.

But is creativity a fundamentally human phenomenon? Or can it be learned by machines?

And if they learn to understand us better than we understand ourselves, could the great AI novel—tailored, of course, to your own predispositions in fiction—be the best you’ll ever read?

Maybe Not a Beach Read
This is the futurist’s view, of course. The reality, as the jury-rigged contraption in Ross Goodwin’s Cadillac for that road trip can attest, is some way off.

“This is very much an imperfect document, a rapid prototyping project. The output isn’t perfect. I don’t think it’s a human novel, or anywhere near it,” Goodwin said of the novel that his machine created. 1 The Road is currently marketed as the first novel written by AI.

Once the neural network has been trained, it can generate any length of text that the author desires, either at random or working from a specific seed word or phrase. Goodwin used the sights and sounds of the road trip to provide these seeds: the novel is written one sentence at a time, based on images, locations, dialogue from the microphone, and even the computer’s own internal clock.

The results are… mixed.

The novel begins suitably enough, quoting the time: “It was nine seventeen in the morning, and the house was heavy.” Descriptions of locations begin according to the Foursquare dataset fed into the algorithm, but rapidly veer off into the weeds, becoming surreal. While experimentation in literature is a wonderful thing, repeatedly quoting longitude and latitude coordinates verbatim is unlikely to win anyone the Booker Prize.

Data In, Art Out?
Neural networks as creative agents have some advantages. They excel at being trained on large datasets, identifying the patterns in those datasets, and producing output that follows those same rules. Music inspired by or written by AI has become a growing subgenre—there’s even a pop album by human-machine collaborators called the Songularity.

A neural network can “listen to” all of Bach and Mozart in hours, and train itself on the works of Shakespeare to produce passable pseudo-Bard. The idea of artificial creativity has become so widespread that there’s even a meme format about forcibly training neural network ‘bots’ on human writing samples, with hilarious consequences—although the best joke was undoubtedly human in origin.

The AI that roamed from New York to New Orleans was an LSTM (long short-term memory) neural net. By default, information contained in individual neurons is preserved, and only small parts can be “forgotten” or “learned” in an individual timestep, rather than neurons being entirely overwritten.

The LSTM architecture performs better than previous recurrent neural networks at tasks such as handwriting and speech recognition. The neural net—and its programmer—looked further in search of literary influences, ingesting 60 million words (360 MB) of raw literature according to Goodwin’s recipe: one third poetry, one third science fiction, and one third “bleak” literature.

In this way, Goodwin has some creative control over the project; the source material influences the machine’s vocabulary and sentence structuring, and hence the tone of the piece.

The Thoughts Beneath the Words
The problem with artificially intelligent novelists is the same problem with conversational artificial intelligence that computer scientists have been trying to solve from Turing’s day. The machines can understand and reproduce complex patterns increasingly better than humans can, but they have no understanding of what these patterns mean.

Goodwin’s neural network spits out sentences one letter at a time, on a tiny printer hooked up to the laptop. Statistical associations such as those tracked by neural nets can form words from letters, and sentences from words, but they know nothing of character or plot.

When talking to a chatbot, the code has no real understanding of what’s been said before, and there is no dataset large enough to train it through all of the billions of possible conversations.

Unless restricted to a predetermined set of options, it loses the thread of the conversation after a reply or two. In a similar way, the creative neural nets have no real grasp of what they’re writing, and no way to produce anything with any overarching coherence or narrative.

Goodwin’s experiment is an attempt to add some coherent backbone to the AI “novel” by repeatedly grounding it with stimuli from the cameras or microphones—the thematic links and narrative provided by the American landscape the neural network drives through.

Goodwin feels that this approach (the car itself moving through the landscape, as if a character) borrows some continuity and coherence from the journey itself. “Coherent prose is the holy grail of natural-language generation—feeling that I had somehow solved a small part of the problem was exhilarating. And I do think it makes a point about language in time that’s unexpected and interesting.”

AI Is Still No Kerouac
A coherent tone and semantic “style” might be enough to produce some vaguely-convincing teenage poetry, as Google did, and experimental fiction that uses neural networks can have intriguing results. But wading through the surreal AI prose of this era, searching for some meaning or motif beyond novelty value, can be a frustrating experience.

Maybe machines can learn the complexities of the human heart and brain, or how to write evocative or entertaining prose. But they’re a long way off, and somehow “more layers!” or a bigger corpus of data doesn’t feel like enough to bridge that gulf.

Real attempts by machines to write fiction have so far been broadly incoherent, but with flashes of poetry—dreamlike, hallucinatory ramblings.

Neural networks might not be capable of writing intricately-plotted works with charm and wit, like Dickens or Dostoevsky, but there’s still an eeriness to trying to decipher the surreal, Finnegans’ Wake mish-mash.

You might see, in the odd line, the flickering ghost of something like consciousness, a deeper understanding. Or you might just see fragments of meaning thrown into a neural network blender, full of hype and fury, obeying rules in an occasionally striking way, but ultimately signifying nothing. In that sense, at least, the RNN’s grappling with metaphor feels like a metaphor for the hype surrounding the latest AI summer as a whole.

Or, as the human author of On The Road put it: “You guys are going somewhere or just going?”

Image Credit: eurobanks / Shutterstock.com Continue reading

Posted in Human Robots

#433796 Creepy AI-Created Portrait Fetches ...

It's the arrival of artificial intelligence in the art world. Continue reading

Posted in Human Robots