Month: February, 2018

When A Machine Learning Algorithm Studied Fine Art Paintings, It Saw Things Art Historians Had Never Noticed


Georges Braque’s Man with a Violin (left) and Pablo Picasso’s Spanish Still Life: Sun and Shadow, both painted in 1912

The task of classifying pieces of fine art is hugely complex. When examining a painting, an art expert can usually determine its style, its genre, the artist and the period to which it belongs.

Art historians often go further by looking for the influences and connections between artists, a task that is even trickier.

So the possibility that a computer might be able to classify paintings and find connections between them at first glance seems laughable.

And yet, that is exactly what Babak Saleh and pals have done at Rutgers University in New Jersey.

These guys have used some of the latest image processing and classifying techniques to automate the process of discovering how great artists have influenced each other.

They have even been able to uncover influences between artists that art historians have never recognized until now.




The way art experts approach this problem is by comparing artworks according to a number of high-level concepts such as the artist’s use of space, texture, form, shape, color and so on.

Experts may also consider the way the artist uses movement in the picture, harmony, variety, balance, contrast, proportion and pattern.

Other important elements can include the subject matter, brushstrokes, meaning, historical context and so on. Clearly, this is a complex business.

So it is easy to imagine that the limited ability computers have for analyzing two-dimensional images would make this process more or less impossible to automate. But Salah and co show how it can be done.

At the heart of their method, is a new technique developed at Dartmouth College in New Hampshire and Microsoft research in Cambridge, UK, for classifying pictures according to the visual concepts that they contain.

These concepts are called classemes and include everything from simple object description such as duck, frisbee, man, wheelbarrow to shades of color to higher-level descriptions such as dead body, body of water, walking and so on.

Comparing images is then a process of comparing the words that describe them, for which there are a number of well-established techniques.

Left: Portrait of Pope Innocent X (1650) by Diego Velazquez. Right: Study After Vel´azquez’s Portrait of Pope Innocent X (1953) by Francis
Bacon

For each painting, they limit the number of concepts and points of interest generated by their method to 3000 in the interests of efficient computation.

This process generates a list of describing words that can be thought of as a kind of vector. The task is then to look for similar vectors using natural language techniques and a machine learning algorithm.

Determining influence is harder though since influence is itself a difficult concept to define. Should one artist be deemed to influence another if one painting has a strong similarity to another?

Or should there be a number of similar paintings and if so how many?

So Saleh and co experiment with a number of different metrics. They end up creating two-dimensional graphs with metrics of different kinds on each axis and then plotting the position of all of the artists in this space to see how they are clustered.

The results are interesting. In many cases, their algorithm clearly identifies influences that art experts have already found.

For example, the graphs show that the Austrian painter Klimt is close to Picasso and Braque and indeed experts are well acquainted with the idea that Klimt was influenced by both these artists.

The algorithm also identifies the influence of the French romantic Delacroix on the French impressionist Bazille, the Norwegian painter Munch’s influence on the German painter Beckmann and Degas’ influence on Caillebotte.

The algorithm is also able to identify individual paintings that have influenced others.

It picked out Georges Braque’s Man with a Violin and Pablo Picasso’s Spanish Still Life: Sun and Shadow, both painted in 1912 with a well-known connection as pictures that helped found the Cubist movement.

And yet a visual inspection shows a clear link. The yellow circles in the images below show similar objects, the red lines show composition and the blue square shows a similar structural element, say Saleh and co.

That is interesting stuff. Of course, Saleh and co do not claim that this kind of algorithm can take the place of an art historian.

After all, the discovery of a link between paintings in this way is just the starting point for further research about an artist’s life and work.

But it is a fascinating insight into the way that machine learning techniques can throw new light on a subject as grand and well studied as the history of art.

Please like, share and tweet this article.

Pass it on: New Scientist

Killer Whales Should Not Be Kept In Captivity

The release of the documentary Blackfish in 2013 shined the national spotlight on the perils and problems of killer whale captivity.

Focusing primarily on killer whale aggression against trainers at SeaWorld and the associated fallout in the wake of veteran trainer Dawn Brancheau’s death in 2010, the film made a strong case against keeping killer whales in captivity.

Killer whales, also known as orcas, aren’t really whales at all. Though in the same infraorder, Cetacea, they are actually dolphins — the largest dolphins, in fact.




They generally weigh between 6,000 and 15,000 pounds when fully grown and stretch 30 feet long from fluke (tail) to rostrum (nose). In the wild, they are apex predators, hunted by no animal apart from man.

In captivity, however, the tides are turned on man. In the half-century that humans have kept orcas in tanks, there have been dozens of documented incidents of aggression, resulting in six deaths.

Orcas form complex societies reminiscent of those seen in chimpanzees. But in captivity, these hierarchies are severely muddled.

Often, a dominant female (orcas are matriarchal) asserts herself, usually violently. In the cramped conditions of captivity, other whales suffer lacerations from her intimidating tooth rakes.

Captive orcas also suffer from a wide range of health problems not seen among wild ones.

 

Their shallow pools render them vulnerable to higher-than-normal levels of ultraviolet radiation from the sun, which suppresses their immune systems.

They also spend more time exposed to the air than in the wild, often moving slowly due to the restrictive size of their habitats. Thus, orcas are prime feeding targets for mosquitos.

In fact, two captive orcas have died from mosquito-borne illnesses. Perhaps the most emblematic health problem associated with captivity is the collapse of the orca’s daunting, shark-like dorsal fin.

Though not actually harmful, the deformity affects over half of killer whales in captivity. Less than 10% of animals in the wild are afflicted.

Killer whales are extremely intelligent creatures, one of the few capable of passing the mirror self-recognition test. It’s no surprise, then, that they are easily subject to boredom in captivity.

One of the ways this manifests is paint nibbling. Whales often use their teeth to peel the paint off of their enclosure’s inner walls, similar to a human biting his fingernails, if his fingernails were made of concrete.

Almost all the whales in SeaWorld wear down their teeth, Hargrove says, dealing damage that requires regular dental procedures to prevent the growth of potentially deadly bacteria.

All of these issues contribute to a startling statistic. While most animals in captivity outlive their wild counterparts, orcas in captivity live shorter. Orca activists claim the gap is wide, while SeaWorld claims that there’s no gap at all.

The best estimates say that captive lifespan is slightly reduced, but improving. Male and female killer whales survive an average of 31 and 46 years in the wild, respectively.

Please like, share and tweet this article.

Pass it on: New Scientist

 

Bitcoin Payments Are Used To Unmask Dark Web Users

Researchers have discovered a way of identifying those who bought or sold goods on the dark web, by forensically connecting them to Bitcoin transactions.

It sounds counter-intuitive. The dark web comprises thousands of hidden services accessed through an anonymity-protecting system, usually Tor.

Bitcoin transactions, meanwhile, are supposed to be pseudonymous, which is to say visible to everyone but not in a way that can easily be connected to someone’s identity.




If you believe that putting these two technologies together should result in perfect anonymity, you might want to read When A Small Leak Sinks A Great Ship to hear some bad news:

Bitcoin lacks retroactive operational security, which means historical pieces of information could be used to identify a certain user.

Which is to say, every Bitcoin transaction that has ever happened exists as a public record, or ledger, that links addresses sending and receiving cryptocurrency.

The task, then, was to find a way to connect these transactions to the online identities of the people responsible for them.

Not easy, you’d assume, but a big weakness of Bitcoin, the dark web, indeed of the whole notion of anonymity on the internet turns out to be the careless way people use social media and specialist forums.

First, researchers trawled 1,500 hidden services on the dark web, from which they managed to uncover 88 active Bitcoin addresses from public data on their landing pages (an address being a single-use token hashed from the owner’s public key).

The same principle was used to uncover 4,100 Bitcoin addresses carelessly advertised on Twitter and 41,000 on the popular BitcoinTalk Forum.

Armed with two sets of Bitcoin addresses – one from the dark web, the second public domain – the researchers set out to connect them, first by using a statistical technique called wallet closure to reliably group lots of transactions to individual wallets.

Thanks to the architecture of the Bitcoin ledger.

This revealed 125 identities that had used dark web services, including WikiLeaks (46 identities), Silk Road (22), the Snowden Defense Fund (11) and The Pirate Bay (10), among others.

De-anonymising these online identities depended on how much information individuals had revealed online but in many cases led to named people in a range of countries.

The number of real people they were able to identify is also incredibly small relative to the volume of Bitcoin transactions heading to or out of the dark web addresses identified.

Nevertheless, given the privacy limitations of Bitcoin, one at least begins to get some sense of what might be driving some dark web users to newer and hypothetically more anonymity-preserving cryptocurrencies such as Monero.

Please like, share and tweet this article.

Pass it on: New Scientist

U.S. Bank Stadium In Minnesota Has Interesting Technology

With construction now complete, U.S. Bank Stadium will soon open its doors.

Those doors might not be what the stadium is most known for, but maybe they should be.

Since the beginning, all the way from design through construction, the stadium’s clear, see-through roof has been the headliner grabbing all the attention.

Which has caused the stadium’s doors to sort of slip under the radar. Which is kind of incredible.

Because on any other stadium, they’d be the star.




Passing through a door like this is like nothing else you’ve ever experienced,” Berg said.

Doors are about transition. You go from a bright sunny day like today, to a dark interior. [But] that’s not what these doors are like. You don’t know whether you’re inside or outside.

“It’s the absolute absence of contrast. That confuses you.

Rotating on hydraulic pistons, they are the five largest pivoting glass doors in the world, 55 feet wide and ranging from 75 to 95 feet high, consisting of 30,000 square feet of glass from Owatonna, attached to door frames manufactured in Tennessee, and weighing, altogether, 40,000 tons.

The Vikings initially thought the NFL might need to create a new policy for their first-in-the-league doors, similar to the one for opening and closing retractable roofs.

But the NFL informed the Vikings in May that its existing retractable roof and wall policy applies: The Vikings must make a decision — doors open or doors closed — at least 90 minutes before kickoff.

As for the factors that will go into that decision? The Vikings are planning to do some testing to determine what happens when the doors are open. Any wind, or other impact?

I think they might have an effect,” Berg said. “Certainly the architects know that there’s a ventilation effect, to having the doors open.

“I suppose there’s a chance that that might affect the game, [for example] the flight of a field goal kick.”

If that ever happens, no doubt the doors would surpass the roof as the U.S. Bank Stadium’s most famous feature.

Well the doors are a little underrated,” Berg said.

Berg’s book is available for pre-order right now from the Minnesota Historical Society Press, or the Vikings website.

Please like, share and tweet this article.

Pass it on: Popular Science

Tom Parr – The 150 Year Old Man

Thomas Parr – also known as Old Tom Parr – lived to the astonishing age of 152. According to legend.

So did he really live to be that old? And if so, what can we learn from him? What’s the key to longevity and living past 100?

Indian Stone Tools Could Dramatically Push Back Date When Modern Humans First Left Africa

We are all children of Africa. As members of the hominin species Homo sapiens, you and I are the product of millions of years of shared evolutionary history of life on Earth.

But as a species we are relatively recent, emerging between 400,000 and 300,000 years ago in East Africa from indigenous archaic populations.

Currently, some of the biggest questions facing palaeoanthropology involve trying to work out how and when early humans left the continent. Was it a single dispersal? Or multiple?

A recent discovery of a jawbone fossil in Israel suggests that there could have been a migration as early as about 180,000 years ago.




But a new study, published in the journal Nature, suggests early humans may have left Africa much earlier than that.

The new research reports the discovery of tools from the Middle Palaeolithic (200,000 to 40,000 years ago) in Tamil Nadu, India.

Surprisingly, the tools date back to 385,000 years ago – which is around the same time as this technology is thought to have first developed by archaic or possibly modern humans in Africa.

This challenges the view, backed by most researchers, that modern humans brought these technologies to India less than 140,000 years ago.

Attirampakkam site

Attirampakkam is located on the banks of a stream of the Kortallaiyar River in northeast Tamil Nadu.

Excavations by a team of Indian researchers revealed abundant layers of stone tools trapped within sediments deposited by streams which ran through the area in prehistory.

The site appears to have been sporadically occupied by apes and early hominins predating Homo sapiens from as far back as 1.7m years ago.

Using a dating technique called infrared-stimulated luminescence – which pinpoints the last time that sediment grains were exposed to light – the authors determined that the silts and gravels which contain the tools date to between 385,000 and 172,000 years ago.

These tools chart the transition from the Acheulean handaxe culture, created by archaic humans of the Lower Palaeolithic, to smaller tools.

The latter were produced by a more sophisticated technique called Levallois – involving the production of stone points and blades.

The tools push the date back for the origins of Middle Palaeolithic technology in India.

Previous studies have suggested that this occurred between 140,000 years and 46,000 years ago, possibly as Homo sapiens migrated into the subcontinent.

But what is perhaps more important, is what these dates mean for the emergence of Homo sapiens and our species’ migrations into the rest of the Old World.

And to understand those implications we need to consider fossils from North Africa and how they are associated with hominin species and technology.

Please like, share and tweet this article.

Pass it on: New Scientist

Could Underwater Sound Waves Be The Key To Early Tsunami Warnings?

Mathematicians think they have devised a way of calculating the size and force of a tsunami in advance of it hitting land, which can help early detection.

Experts say naturally occurring high-speed acoustic gravity waves are created after “tsunami trigger events”.

Cardiff University scientists hope to make a real-time early warning system.

Alaska was under a tsunami warning earlier this week after a 7.9-magnitude earthquake struck 280km (173 miles) off the coast of the American state.

The deadliest recorded tsunami was the 2004 Boxing Day Indian Ocean tsunami, which killed almost 230,000 people in 11 different countries.




But scientists in Cardiff hope to help give extra warning time for tsunamis by using the fast-moving underwater sound waves.

By taking measurements of acoustic gravity waves, we basically have everything we need to set off a tsunami alarm,” said Dr Usama Kadri, lead author for the study from Cardiff University’s school of mathematics.

Underwater earthquakes are triggered by the movement of tectonic plates on the ocean floor and are the main cause of tsunamis.

Scientists say sound waves can travel over 10 times faster than tsunamis and spread out in all directions, regardless of the trajectory of the tsunami, making them easy to pick up using standard underwater hydrophones.

They say this is an ideal source of information for early warning systems.

In a new study published in the Journal of Fluid Mechanics, Cardiff University scientists show how the key characteristics of an earthquake – such as its location, duration, dimensions, orientation and speed – can be determined when the gravity waves are detected by a single hydrophone in the ocean.

The sound waves move through the deep ocean at the speed of sound and can travel thousands of meters below the surface.

Tsunamis are currently detected by floating buoys that are able to measure pressure changes in the ocean caused by tsunamis.

However, experts say the technology relies on a tsunami physically reaching the buoys.

The current technology also requires the distribution of a huge number of expensive buoys in oceans all around the world.

Though we can currently measure earthquakes using seismic sensors, these do not tell us if tsunamis are likely to follow,” Dr Kadri continued.

Using sound signals in the water, we can identify the characteristics of the earthquake fault, from which we can then calculate the characteristics of a tsunami. Since our solution is analytical, everything can be calculated in near real-time.

Our aim is to be able to set off a tsunami alarm within a few minutes from recording the sound signals in a hydrophone station.

Please like, share and tweet this article.

Pass it on: New Scientist

What Is Alphabet’s Chronicle?

Google’s parent company Alphabet has launched a business that will specialize in leveraging machine learning in cyber security, called Chronicle.

Chief executive Stephen Gillett says that the company will be split in two. On the one hand it will provide a cyber security and analytics platform that will target enterprise customers and help them “better manage and understand their own security-related data“.

The other side of the business will be VirusTotal, which is a malware intelligence service Google picked up in 2012. This will continue to operate as it has been doing.




For some years now a slew of security vendors have touted machine learning as their key differentiator against rivals on the market. There is an aspect of snake oil to some of it – see our market analysis here.

But there are also companies like the darlings of the infosec market at the moment, Darktrace, that are using genuine machine learning for threat detection.

It’s no secret that Alphabet and Google are at the frontier of machine learning and artificial intelligence.

Writing in Medium, Chronicle CEO Stephen Gillett says that while the company will be an independent business under the Google umbrella, it will “have the benefit of being able to consult the world-class experts in machine learning and cloud computing” at Alphabet.

Where did Chronicle come from?

Chronicle emerged from the labs of Alphabet’s mysteriously named X – Google’s incubation hub for ‘moonshot’ projects, in 2016, plus VirusTotal which Google bought in 2012.

CEO Stephen Gillet began working at Google in 2015 and has a history of work in cyber security companies.

Other people in leadership roles at Chronicle include Mike Wiacek and Shapor Naghibzadeh, who together have more than 20 years of security experience at Google.

Bernardo Quintero of VirusTotal will continue to work with Chronicle.

Please like, share and tweet this article.

Pass it on: Popular Science

Is Butter Really A Carb?

Turns out, one of the most famous lines from our favorite chick flick, Mean Girls, is wrong.

Regina George is trying to lose weight and cutting out carbs, and asks Cady (Lindsey Lohan’s character) if butter is a carb.




Cady famously replies with a condescending “yes”, but it turns out, butter is in actual fact not a carb. Okay, well it is, but it’s a very low carb. Low as in it contains only  0,1g of carbs.

Butter is used in low carb diets like banting. So technically, Regina could have – and should have – had that butter instead of the cheese fries she decided to get instead.

Please like, share and tweet this article.

Pass it on: Popular Science