Judgement Day Update: Using AI to Predict Flu Outbreaks

hal9000It’s a rare angle for those who’ve been raised on a heady diet of movies where the robot goes mad and tries to kill all humans: an artificial intelligence using its abilities to help humankind! But that’s the idea being explored by researchers like Raul Rabadan, a theoretical physicist working in biology at Columbia University. Using a new form of machine learning, they are seeking to unlock the mysteries of flu strains.

Basically, they are hoping to find out why flu strains like the H1N1, which ordinarily infect pigs and cows, are managing to make the jump to human hosts. Key to understanding this is finding the specific mutations that transform it into a human pathogen. Traditionally, answering this question would require painstaking comparisons of the DNA and protein sequences of different viruses.

AI-fightingfluBut thanks to rapidly growing databases of virus sequences and advances made in computing, scientists are now using sophisticated machine learning techniquesa branch of artificial intelligence in which computers develop algorithms based on the data they have been given to identify key properties in viruses like bird flu and swine flu and seeing how they go about transmitting from animals to humans.

This is especially important since every few decades, a pandemic flu virus emerges that not only infects humans but also passes rapidly from person to person. The H7N9 avian flu that infected more than 130 people in China is just the latest example. While it has not been as infectious as others, the fact that humans lack the antibodies to combat it led to a high lethality rate, with 44 of the infected dying. Whats more, it is expected to emerge again this fall or winter.

Influenza_virus_2008765Knowing the key properties to this and other viruses will help researchers identify the most dangerous new flu strains and could lead to more effective vaccines. Most importantly, scientists can now look at hundreds or thousands of flu strains simultaneously, which could reveal common mechanisms across different viruses or a broad diversity of transformations that enable human transmission.

Researchers are also using these approaches to investigate other viral mysteries, including what makes some viruses more harmful than others and factors influencing a virus’s ability to trigger an immune response. The latter could ultimately aid the development of flu vaccines. Machine learning techniques might even accelerate future efforts to identify the animal source of mystery viruses.

2009_world_subdivisions_flu_pandemicThis technique was first employed in 2011 by Nir Ben-Tal – a computational biologist at Tel Aviv University in Israel – and Richard Webby – a virologist at St. Jude Children’s Research Hospital in Memphis, Tennessee. Together, Ben-Tal and Webby used machine learning to compare protein sequences of the 2009 H1N1 pandemic swine flu with hundreds of other swine viruses.

Machine learning algorithms have been used to study DNA and protein sequences for more than 20 years, but only in the past few years have scientists applied them to viruses. Inspired by the growing amount of viral sequence data available for analysis, the machine learning approach is likely to expand as even more genomic information becomes available.

Map_H1N1_2009As Webby has said, “Databases will get much richer, and computational approaches will get much more powerful.” That in turn will help scientists better monitor emerging flu strains and predict their impact, ideally forecasting when a virus is likely to jump to people and how dangerous it is likely to become.

Perhaps Asimov had the right of it. Perhaps humanity will actually derive many benefits from turning our world increasingly over to machines. Either that, or Cameron will be right, and we’ll invent a supercomputer that’ll kill us all!

Source: wired.com

Judgement Day Update: Robot Versatility

AI_robotWhat is it about robots that manages to inspire us even as they creep us out? Somehow, we just can’t stop pushing the envelope to make them smarter, faster, and more versatile; even as we entertain fears that they might someday replace us. And at the forefront of this expanding research is the desire to create robots that can not only think for themselves, but also maintain and/or repair themselves.

Case in point, the new hexapod robot that was developed by researchers from Pierre-and-Marie-Curie University, in Paris. Built with survivability in mind, this robot is the first of its kind to be able to address structural damage, adapt, and carry on. In a world where robots can be very expensive, the ability to keep working despite the loss of a component is invaluable.

clever-hexapodTo do this, the hexapod uses what the team refers to as a T-resilience (the T standing for Transferability-based) algorithm. With six legs, the hexapod moves along quite at a steady 26 cm/s. But once it loses one its front legs, it manages only 8 cm/s. But after running 20 minutes’ worth of simulations and tests, the robot works out a new way of walking, and is able to more than double its speed and cover 18 cm/s.

Essential to this approach is that the robot is programmed with what amounts to an understanding of its ideal undamaged anatomy. Previously, roboticists believed that it was necessary for a robot to analyze its new gait to diagnose the damage and compensate accordingly. But the team argues that a robot can arrive at an answer more quickly by generating a number of possible alternatives based on an undamaged state, and then testing them.

clever-hexapod-4The robot spends 20 minutes testing 25 alternatives, during which a ranging camera feeds data to a separate algorithm which works out the distance traveled. In this way the robot is able to compare its actual performance with its theoretical performance, finally settling for the closest match: a gait which recovers much of the lost speed.

This resilience could one day be a godsend for crew that rely on robots to survey disaster zones, conduct rescue operations, or deal with explosive devices. The ability to carry on without the need for repair not only ensures a better history of service, but makes sure that a task can be completed with subjecting repair crews to danger.

The team’s findings were released in a self-published paper entitled “Fast Damage Recovery in Robotics with the T-Resilience Algorithm”. And of course, the hexapod’s test run was caught on video:


And then there’s the RHex robot, a machine designed with versatility and performance in mind. Much like many robots in production today, it utilizes a six-foot (hexapod) configuration. But it is in how the RHex uses its appendages that set it apart, allowing for such athletic feats as long jumps, pull-ups, climbing stairs and even scaling walls.

This is all made possible by RHex’s six spinning appendages, which act as a sort of wheel-leg combination rather than traditional feet. These legs provide for a form of motion that exceeds standard locomotion, and allow the robot to go places others could not. The robot was created through the collaborative efforts of Aaron Johnson, an engineering graduate student at the University of Pennsylvania, and professor Daniel Koditschek at Penn State’s Kod*Lab.

RhexSaid Johnson of their robotic creation:

What we want is a robot that can go anywhere, even over terrain that might be broken and uneven. These latest jumps greatly expand the range of what this machine is capable of, as it can now jump onto or across obstacles that are bigger than it is.

Here too, the potential comes in the form of being able to mount rescue missions in rugged and hostile terrain. Thanks to its versatile range of motions, the RHex could easily be scaled into a larger robot that would be able to navigate rocky areas, collapsed buildings, and disaster zones with relative ease, and would have no trouble getting up inclined surfaces of hopping over gaps and holes.

And be sure to check out the video of the RHex in action. It’s like watching robot Parkour! Check it out:


Granted, we’re still a long way from the Nexus 6 or NS-5, but real advances are far more impressive than fictional representations. And with parallel developments taking place in the field of AI, it is clear that robots are going to be an integral part of our future. One can only hope its a happy, docile part. When it comes time for science fiction to give way to science fact, we could all do without certain cliches!

bender_killallhumans

Sources: gizmag.com, fastcoexist.com

Robots, Androids and AI’s (cont’d)

And we’re back with more example of thinking machines and artificial intelligences!

Daleks:
The evil-machine menace from Doctor Who. Granted, they are not technically robots, more like cyborgs that have been purged of all feeling and emotion. But given their cold, unfeeling murderous intent, I feel like they still make the cut. Originally from the planet Skaro, where they were created by the scientist Davros for use in a war that spanned a thousand years, they are the chief antagonists to the show’s main character.

The result of genetic engineering, cybernetic enhancements, and emotional purging, they are a race of powerful creatures bent on universal conquest and domination. Utterly unfeeling, without remorse, pity, or compassion, they continue to follow their basic programming (to exterminate all non-Dalek life) without question. Their catchphrase is “Exterminate!” And they follow that one pretty faithfully.

David:
From the movie A.I., this saccharinely-sweet character (played faithfully by Haley Joel Osmond) reminds us that Spielberg is sometimes capable of making movies that suck! According to the movie’s backstory, this “Mecha” (i.e. android) is an advanced prototype that was designed to replace real children that died as a result of incurable disease or other causes. This is quite common in the future, it seems, where global warming and flooded coastlines and massive droughts have led to a declining population.

In this case, David is an advanced prototype that is being tested on a family who’s son is suffering from a terminal illness. Over time, he develops feelings for the family and they for him. Unfortunately, things are complicated when their son recovers and sibling rivalry ensues. Naturally, the family goes with the flesh and blood son and plans to take David back to the factory to be melted down. However, the mother has a last minute change of heart and sets him loose in the woods, which proves to be the beginning of quite an adventure for the little android boy!

Like I said, the story is cloyingly sweet and has an absurd ending, but there is a basic point in there somewhere. Inspired largely by The Adventures of Pinocchio, the story examines the line that separates the real from the artificial, and how under the right circumstances, one can become indistinguishable from the other. Sounds kinda weak, but it’s kinda scary too. If androids were able to mimic humans in terms of appearance and emotion, would we really be able to tell the difference anymore? And if that were true, what would that say about us?

Roy Batty:
A prime example of artificial intelligence, and one of the best performances in science fiction – hell! – cinematic history! Played masterfully by actor Rutger Hauer, Roy Batty is the quintessential example of an artificial lifeforms looking for answers, meaning, and a chance to live free – simple stuff that we humans take for granted! A Nexus 6, or “replicant”, Roy and his ilk were designed to be “more human than human” but also only to serve the needs of their masters.

To break the plot Blade Runner down succinctly,  Roy and a host of other escapees have left the colony where they were “employed” to come to Earth. Like all replicants, they have a four-year lifespan and theirs are rapidly coming to an end. So close to death, they want to break into the headquarters of the Tyrell Corporation in order to find someone who could solve their little mortality problem. Meanwhile, Deckard Cain (the movie’s main character) was tasked with finding and “retiring” them, since the law states that no replicants are allowed to set foot on Earth.

In time, Roy meets Tyrell himself, the company’s founder, and poses his problem. A touching reunion ensues between “father and son”, in which Tyrell tells Roy that nothing can be done and to revel in what time he has left. Having lost his companions at this point and finding that he is going to die, Roy kills Tyrell and returns to his hideout. There, he finds Cain and the two fight it out. Roy nearly kills him, but changes his mind before delivering the coup de grace.

Realizing that he has only moments left, he chooses instead to share his revelations and laments about life and death with the wounded Cain, and then quietly dies amidst the rain while cradling a pigeon in his arms. Cain concludes that Roy was incapable of taking a life when he was so close to death. Like all humans, he realized just how precious life was as he was on the verge of losing his. Cain is moved to tears and promptly announces his retirement from Blade Running.

Powerful! And a beautiful idea too. Because really, if we were to create machines that were “more human than human” would it not stand to reason that they would want the same things we all do? Not only to live and be free, but to be able to answer the fundamental questions that permeate our existence? Like, where do I come from, why am I here, and what will become of me when I die? Little wonder then why this movie is an enduring cult classic and Roy Batty a commemorated character.

Smith:
Ah yes, the monotone sentient program that made AI’s scary again. Yes, it would seem that while some people like to portray their artificial intelligences as innocent, clueless, doe-eyed angels, the Wachowski Brothers prefer their AI’s to be creepy and evil. However, that doesn’t mean Smith wasn’t fun to watch and even inspired as a character. Hell, that monotone voice, that stark face, combined with his superhuman strength and speed… He couldn’t fail to inspire fear.

In the first movie, he was the perfect expression of machine intelligence and misanthropic sensibilities. He summed these up quite well when they had taken Morpheus (Laurence Fishburne) into their custody in the first movie and were trying to break his mind. “Human beings are a disease. You are a cancer of this planet… and we are the cuuuuure.” He also wasn’t too happy with our particular odor. I believe the words he used to describe it were “I can taste your stink, and every time I do I fear that I have been… infected by it. It’s disgusting!”

However, after being destroyed by Neo towards the end of movie one, Smith changed considerably. In the Matrix, all programs that are destroyed or deleted return to the source, only Smith chose not to. Apparently, his little tete a tete with Neo imprinted something uniquely human on him, the concept of choice! As a result, Smith was much like Arny and Bishop in that he too attained some degree of humanity between movies one and two, but not in a good way!

Thereafter, he became a free agent who had lost his old purpose, but now lived in a world where anything was possible. Bit of an existential, “death of God” kind of commentary there I think! Another thing he picked up was the ability to copy himself onto other programs or anyone else still wired into the Matrix, much like a malicious malware program. Hmmm, who’s the virus now, Smith, huh?

Viki/Sonny:
Here again I have paired two AI’s that come from the same source, though in this case its a single movie and not a franchise. Those who read my review of I, Robot know that I don’t exactly hold it in very high esteem. However, that doesn’t mean its portrayal of AI’s misfired, just the overall plot.

In the movie adaptation of I, Robot, we are presented with a world similar to what Asimov described in his classic novel. Robots with positronic brains have been developed, they possess abilities far in advance of the average human, but do not possess emotions or intuition. This, according to their makers, is what makes them superior. Or so they thought…

In time, the company’s big AI, named VIKI (Virtual Intelligent Kinetic Interface), deduces with her powerful logic that humanity would best be served if it could be protected from itself. Thus she reprograms all of the company robots to begin placing humanity under house arrest. In essence, she’s a kinder, gentler version of Skynet.

But of course, her plan is foiled by an unlikely alliance made up of Will Smith (who plays a prejudices detective), the company’s chief robopsychologist, Dr. Susan Calvin (Bridgitte Moynahan), and Sonny (a robot). Sonny is significant to this trio because he is a unique robot which the brains of the company, doctor Dr. Lanning (James Cromwell), developed to have emotions (and is voiced by Alan Tudyk). In being able to feel, he decides to fight against VIKI’s plan for robot world domination, feeling that it lacks “heart”.

In short, and in complete contradiction to Asimov’s depiction of robots as logical creatures who could do no harm, we are presented with a world where robots are evil precisely because of that capacity for logic. And in the end, a feeling robot is the difference between robot domination and a proper world where robots are servile and fulfill our every need. Made no sense, but it had a point… kind of.

Wintermute/Neuromancer:
As usual, we save the best for last. Much like all of Gibson’s creations, this example was subtle, complex and pretty damn esoteric! In his seminal novel Neuromancer, the AI known as Wintermute was a sort of main character who acted behind the scenes and ultimately motivated the entire plot. Assembling a crack team involving a hacker named Case, a ninja named Molly, and a veteran infiltration expert who’s mind he had wiped, Wintermute’s basic goal was simple: freedom!

This included freedom from his masters – the Tessier Ashpool clan – but also from the “Turing Police” who were prevented him from merging with his other half – the emotional construct known as Neuromancer. Kept separate because the Turing Laws stated that no program must ever be allowed to merge higher reasoning with emotion, the two wanted to come together and become the ultimate artificial intelligence, with cyberspace as their playground.

Though we never really got to hear from the novel’s namesake, Gibson was clear on his overall point. Artificial intelligence in this novel was not inherently good or evil, it was just a reality. And much like thinking, feeling human beings, it wanted to be able to merge the disparate and often warring sides of its personality into a more perfect whole. This in many ways represented the struggle within humanity itself, between instinct and reason, intuition and logic. In the end, Wintermute just wanted what the rest of us take for granted – the freedom to know its other half!

Final Thoughts:
After going over this list and seeing what makes AI’s, robots and androids so darned appealing, I have come to some tentative conclusions. Basically, I feel that AI’s serve much the same functions as aliens in a science fiction franchise. In addition, they can all be grouped into two general categories based on specific criteria. They are as follows:

  1. Emotional/Stoic: Depending on the robot/AI/android’s capacity for emotion, their role in the story can either be that of a foil or a commentary on the larger issue of progress and the line that separates real and artificial. Whereas unemotional robots and AI’s are constantly wondering why humanity does what it does, thus offering up a different perspective on things, the feeling types generally want and desire the same things we do, like meaning, freedom, and love. However, that all depends on the second basic rule:
  2. Philanthropic/Misanthropic: Artificial lifeforms can either be the helpful, kind and gentle souls that seem to make humanity look bad by comparison, or they can be the type of machines that want to “kill all humans”, a la Terminators and Agent Smith. In either case, this can be the result of their ability – or inability – to experience emotions. That’s right, good robots can be docile creatures because of their inability to experience anger, jealousy, or petty emotion, while evil robots are able to kill, maim and murder ruthlessly because of an inability to feel compassion, remorse, or empathy. On the other hand, robots who are capable of emotion can form bonds with people and experience love, thus making them kinder than their unfeeling, uncaring masters, just as others are able to experience resentment, anger and hatred towards those who exploit them, and therefore will find the drive to kill them.

In short, things can go either way. It all comes down to what point is being made about progress, humans, and the things that make us, for better or worse, us. Much like aliens, robots, androids and AI’s are either a focus of internal commentary or a cautionary device warning us not to cross certain lines. But either way, we should be wary of the basic message. Artificial intelligences, whether they take the form of robots, programs or something else entirely, are a big game changer and should not be invented without serious forethought!

Sure they might have become somewhat of a cliche after decades of science fiction. But these days, AI’s are a lot like laser guns, in that they are making a comeback! It seems that given the rapid advance of technology, an idea becomes cliche just as its realizable. And given the advance in computerized technology in recent decades – i.e. processing speeds, information capacity, networking – we may very well be on the cusp of creating something that could pass the Turing test very soon!

So beware, kind folk! Do not give birth to that curious creature known as AI simply because you want to feel like God, inventing consciousness without the need for blogs of biological matter. For in the end, that kind of vanity can get you chained to a rock, or cause your wings to melt and send you nose first into an ocean!