Tag: University of Illinois

Judgement Day Update: Deep Learning and AIs

brain_chip2Moore’s Law, a sort of general law of computing that was coined by Intel co-founder Gordon E. Moore, states that the number of transistors on integrated circuits doubles approximately every two years. In accordance with this observation, which was made back in 1965, computers have evolved to the point where they are over 650 times faster than they were in the early 1970’s.

Given this trend, scientists and futurists have long suspected that it is only a matter of time before computers surpass human beings in terms of raw intelligence. Already, supercomputers like IBM’s Watson, University of Illinois’ “Blue Waters”, and MIT’s ConceptNet 4 have demonstrated that they are superior in terms of raw knowledge and retention. But when it comes to actual thinking – i.e. independent reasoning and common sense – they are seriously lacking.

IBM_Blue_Gene_P_supercomputerYes, despite all the progress made in the field of computing in the past 40 years, bridging that gap between computational power and artificial intelligence has remained elusive. And at this point, no one is quite sure what it will take to make that leap happen. However, a number of developers and designers are coming up with breakthroughs that may constitute a small jump.

One such breakthrough comes from Standford University, where computer engineers have come up with a new algorithm that could give computers the power to more reliably interpret language. Called Neural Analysis of Sentiment – or NaSent for short – the algorithm seeks to improve on current methods of written language analysis by drawing inspiration from the human brain.

AI_picNaSent is part of a movement in computer science known as deep learning, a new field that seeks to build programs that can process data in much the same way the brain does. According to Richard Socher, the Stanford University grad student who helped develop NaSent:

In the past, sentiment analysis has largely focused on models that ignore word order or rely on human experts. While this works for really simple examples, it will never reach human-level understanding because word meaning changes in context and even experts cannot accurately define all the subtleties of how sentiment works. Our deep learning model solves both problems.

Socher was joined by artificial-intelligence researchers Chris Manning, a Professor of Computer Science and Linguistics at Standford, and Andrew Ng, one of the engineers behind Google’s deep learning project and who is working on creating the “Google Brain”.  Their aim is to develop algorithms that can operate without continued help from humans.

word_cloudWhat sets this algorithm apart is its ability to identify the meaning of words in the context of phrases and sentences. This is a big change from previous methods of sentiment analysis which have been limited to parsing through a collection of words and ranking them as either positive or negative without taking word order into account.

To create NaSent, Socher and his team used 12,000 sentences taken from the movie reviews website Rotten Tomatoes. They split these sentences into roughly 214,000 phrases that were labeled as very negative, negative, neutral, positive, or very positive, and then they fed this labeled data into the system, which NaSent then used to predict whether sentences were positive, neutral or negative on its own.

metadataAccording to the researchers, NaSent is about 85 percent accurate, a 5% improvement over previous models. However, Socher and his team are working on increasing that rate of success by feeding the system more data from Twitter and the Internet Movie Database, and have also set up a live demo where people can type in their own sentences.

And whereas the field of deep learning began in the academic world, it has since spread to web giants such as Google and Facebook. In both cases, the companies have taken to hiring people familiar with the field to help them sort through the growing mountains of data they are indexing to help them improve their products, particularly those that already rely on machine learning.

In Google’s case, these include voice recognition technology, as well as their ongoing work in neural networks. In the case of Facebook, they are looking to deep learning to help them improve Graph Search, a search engine that allows people to search activity on their network. This presents a challenge, mainly because machines have a hard time interpreting the nuances of human language. Just ask Watson, IBM’s own supercomputer!

Sources: wired, (2), (3), engineering.stanford.edu

Judgement Day Update: A.I. Equivalent to Four Year Old Mind

artificial_intelligence1Ever since computers were first invented, scientists and futurists have dreamed of the day when computers might be capable of autonomous reasoning and be able to surpass human beings. In the past few decades, it has become apparent that simply throwing more processing power at the problem of true artificial intelligence isn’t enough. The human brain remains several orders more complex than the typical AI, but researchers are getting closer.

One such effort is ConceptNet 4, a semantic network being developed by MIT. This AI system contains a large store of information that is used to teach the system about various concepts. But more importantly, it is designed to process the relationship between things. Much like the Google Neural Net, it is designed to learn and grow to the point that it will be able to reason autonomously.

child-ai-brainRecently, researchers at the University of Illinois at Chicago decided to put the ConceptNet through an IQ test. To do this, they used the Wechsler Preschool and Primary Scale of Intelligence Test, which is one of the common assessments used on small children. ConceptNet passed the test, scoring on par with a four-year-old in overall IQ. However, the team points out it would be worrisome to find a real child with lopsided scores like those received by the AI.

The system performed above average on parts of the test that have to do with vocabulary and recognizing the similarities between two items. However, the computer did significantly worse on the comprehension questions, which test a little one’s ability to understand practical concepts based on learned information. In short, the computer showed relational reasoning, but was lacking in common sense.

Neuromorphic-chip-640x353This is the missing piece of the puzzle for ConceptNet and those like it. An artificial intelligence like this one might have access to a lot of data, but it can’t draw on it to make rational judgements. ConceptNet might know that water freezes at 32 degrees, but it doesn’t know how to get from that concept to the idea that ice is cold. This is basically common sense — humans (even children) have it and computers don’t.

There’s no easy way to fabricate implicit information and common sense into an AI system and so far, no known machine has shown the ability. Even IBM’s Watson trivia computer isn’t capable of showing basic common sense, and though multiple solutions have been proposed – from neuromorphic chips to biomimetic circuitry – nothing is bearing fruit just yet.

AIBut of course, the MIT research team is already hard at work on ConceptNet 5, a more sophisticated neural net computer that is open source and available on GitHub. But for the time being, its clear that a machine will be restricted to processing information and incapable of making basic decisions. Good thing too! The sooner they can think for themselves, the sooner they can decide we’re in their way!

Source: extremetech.com

The Future of Medicine: Smartphone Medicine!

iphone_specIt’s no secret that the exponential growth in smartphone use has been paralleled by a similar growth in what they can do. Everyday, new and interesting apps are developed which give people the ability to access new kinds of information, interface with other devices, and even perform a range of scans on themselves. It is this latter two aspect of development which is especially exciting, as it is opening the door to medical applications.

Yes, in addition to temporary tattoos and tiny medimachines that can be monitored from your smartphone or other mobile computing device, there is also a range of apps that allow you to test your eyesight and even conduct ultrasounds on yourself. But perhaps most impressive is the new Smartphone Spectrometer, an iPhone program which will allow users to diagnose their own illnesses.

iphone_spec2Consisting of an iPhone cradle, phone and app, this spectrometer costs just $200 and has the same level of diagnostic accuracy as a $50,000 machine, according to Brian Cunningham, a professor at the University of Illinois, who developed it with his students. Using the phone’s camera and a series of optical components in the cradle, the machine detects the light spectrum passing through a liquid sample.

This liquid can consist of urine or blood, any of the body’s natural fluids that are exhibit traces of harmful infection when they are picked up by the body. By comparing the sample’s spectrum to spectrums for target molecules, such as toxins or bacteria, it’s possible to work out how much is in the sample. In short, a quickie diagnosis for the cost of a fancy new phone.

Granted there are limitations at this point. For one, the device is nowhere near as efficient as its industrial counterpart. Whereas automated $50,000 version can process up to 100 samples at a time, the iPhone spectrometer can only do one at a time. But by the time Cunningham and his team plan on commercializing the design, they hope to increase that efficiency by a few magnitudes.

iphone_spec1On the plus side, the device is far more portable than any other known spectrometer. Whereas a lab is fixed in place and has to process thousands of samples at any given time, leading to waiting lists, this device can be used just about anywhere. In addition, there’s no loss of accuracy. As Cunningham explained:

We were using the same kits you can use to detect cancer markers, HIV infections, or certain toxins, putting the liquid into our cartridge and measuring it on the phone. We have compared the measurements from full pieces of equipment, and we get the same outcome.

Cunningham is currently filing a patent application and looking for investment. He also has a grant from the National Science Foundation to develop an Android version. And while he doesn’t think smartphone-based devices will replace standard spectrometry machines with long track records, and F.D.A approval, he does believe they could enable more testing.

publiclaboratoryThis is especially in countries where government-regulated testing is harder to come by, or where medical facilities are under-supplied or waiting lists are prohibitively long. With diseases like cancer and HIV, early detection can be the difference between life and death, which is a major advantage, according to Cunningham:

In the future, it’ll be possible for someone to monitor themselves without having to go to a hospital. For example, that might be monitoring their cardiac disease or cancer treatment. They could do a simple test at home every day, and all that information could be monitored by their physician without them having to go in.

But of course, the new iPhone is not alone. Many other variations are coming out, such as the PublicLaboratory Mobile Spectrometer, or Androids own version of the Spectral Workbench. And of course, this all calls to mind the miniature spectrometer that Jack Andraka, the 16-year old who invented a low-cost litmus test for pancreatic cancer and who won the 2012 Intel International Science and Engineering Fair (ISEF). That’s him in the middle of the picture below:

ISEF2012-Top-Three-WinnersIt’s the age of mobile medicine, my friends. Thanks to miniaturization, nanofabrication, wireless technology, mobile devices, and an almost daily rate of improvement in medical technology, we are entering into an age where early detection and cost-saving devices are making medicine more affordable and accessible.

In addition, all this progress is likely to add up to many lives being saved, especially in developing regions or low-income communities. It’s always encouraging when technological advances have the effect of narrowing the gap between the haves and the have nots, rather than widening it.

And of course, there’s a video of the smartphone spectrometer at work, courtesy of Cunningham’s research team and the University of Illinois:


Source: fast.coexist.com

Supercomputer Creates Atomic Model of HIV

DNA-1The ongoing fight to end HIV has been a long and arduous one, but progress is being made. In addition to potential treatments being created that have shown promise, there are also efforts being mounted to understand how the virus works at an atomic level. This is great news, for as any practitioner of medicine will tell you, understanding a disease and knowing how to strike at the heart of it is the key to stopping it and making sure future generations don’t have to fear it.

In recent years, several major breakthroughs were announced for the treatment of HIV, treatments which many heralded as cures. In January of last year, the Danish Research Council awarded funding to a group of researchers who demonstrated that HIV could be “flushed” from infected cells where it tends to congregate and protect itself. Combined with vaccinations that turbocharge the body’s immune system, this method proved effective at eliminating the HIV virus in infected cells.

HIV-budding-ColorAnother came back in November, when researchers at Caltech were even able to successfully spawn a significant amount of HIV antibodies in lab mice by using a new approach, known as Vectored ImmunoProphylaxis (VIP). An inversion of the traditional vaccination method, this new method produced plenty of HIV-preventing antibodies which they believed could be fashioned into a  vaccine.

And finally, there were the experiments being conducted over at the Washington University School of Medicine, where researchers designed a solution that employed bee venom and a nanoparticle delivery system. Knowing that bee venom is capable of killing HIV, and that the virus is thousands of times smaller than your average cell, the solution proved quite effective at filtering out the virus and killing it while leaving surrounding tissue unharmed. Taken together, these two proposed solutions have left many thinking a cure is just around the corner.

blue-waters-super-computer-at-petascale-020908Nevertheless, in order for this virus to truly be beaten, we need to understand it better. Hence why a group of scientists – using the University of Illinois’ “Blue Waters” supercomputer — have developed a new series of computer models that are finally giving researchers an atomic-level look at the formidable barrier mechanism enclosing the heart of the virus.

For example, its been known for some time that the HIV virus it’s covered in several layers of protective proteins. But beneath that outer shell resides a conical structure called the capsid, which houses the virus’ payload of genetic material. (See diagram below.) When HIV invades a cell, it’s the capsid that opens up to initiate the takeover process, allowing the virus to replicate inside the healthy host cell. Better understanding of how this mysterious delivery system operates could be one of the final steps to finding a cure.

HIVAnd that’s where the modelling software really comes into play. How and when the HIV cell opens to deliver the capsid has long eluded researchers, and as Klaus Schulten, a physicist that was part of the team that modeled the virus, pointed out: “The timing of the opening of the capsid is essential for the degree of virulence of the virus.”

Using the Blue Waters, Schulten and his associates managed to map out the model all 64 million of the capsid’s atoms. Through countless simulations, they also discovered that the capsid’s microscopic outer casing is composed of 216 hexagon-shaped proteins that fit together in a honeycomb formation. These hexagonal structures are what give the capsid it’s tough outer shell and allow it to be such a harmful and insidious killer.

AIDS_memorialThis painstakingly delicate process would have been unthinkable until just a few years ago, and represents the most complete picture of the HIV virus to date. What’s more, knowing what HIV looks like at the atomic level will help scientists to understand the timing of the virus’ delivery system. Since the opening of the virus’ protective layer is when it’s most vulnerable, Schulten and his colleagues hope to determine the precise timing of this event so a treatment can be developed that could attacks the virus at this exact moment.

Think of it as throwing a bomb into the mouth of a terrible war machine, right as it opens up its armored maw to bite you! Better yet, think of it as another step on the road to ending one of the greatest plagues humankind has ever had to deal with. Safety for the future, and justice for the victims!

Sources: popularscience.com, theweek.com, (2)

The Future is Here: Batteries for Stretchable Implants

Stretchable-battery1One of the newest and greatest developments in medical technology of late has been the creation of electronics that can stretch and flex. Increasingly, scientists are developing flexible electronics like video displays and solar panels that could make their way into clothing or even bodies. But of course, some challenges remain, specifically in how to power these devices.

Thus far, researchers have been able to develop batteries that are thin and bendable, flexibility has proven more of a challenge. In addition, no stretchable batteries have thus far offered rechargeability with high the kind of storage capacity that one might expect from the lithium-ion technology now powering many smartphones, tablets, laptops and other mobile devices.

flexbatteryHowever, that may be changing thanks to two research scientists – Yonggang Huang from Northwestern University and John A. Rogers University of Illinois. Together, they have unveiled a rechargeable lithium-ion battery that can be stretched, twisted and bended, and is still capable of powering electronics. What’s more, the power and voltage of this battery are similar to a conventional lithium-ion battery and can be used anywhere, including the inside of the human body.

Whereas previous batteries of its type had a hard time stretching up to 100 percent of their original size, this new design is capable of stretching up to 300 percent. Huang and Rogers have indicated that this will make it ideal for powering implantable electronics that are designed for monitoring brain waves or heart activity. What’s more, it can be recharged wirelessly and has been tested up to 20 cycles of recharging with little loss in capacity.

Stretchable-batteryFor their stretchable electronic circuits, the two developed an array of tiny circuit elements connected by metal wire “pop-up bridges.” Typically, this approach works for circuits but not for a stretchable battery, where components must be packed tightly to produce a powerful enough current. Huang’s design solution is to use metal wire interconnects that are long, wavy lines, filling the small space between battery components.

In a paper published on Feb. 26, 2013 in the online journal Nature Communications, Huang described the process of creating their new design:

“We start with a lot of battery components side by side in a very small space, and we connect them with tightly packed, long wavy lines. These wires provide the flexibility. When we stretch the battery, the wavy interconnecting lines unfurl, much like yarn unspooling. And we can stretch the device a great deal and still have a working battery.”

No telling when the first stretchable electronic implant will be available for commercial use, but now that we have the battery issue worked out, its only a matter of time before hospitals and patient care services are placing them in patients to monitor their health and vitals. Combined with the latest in personal computing and wireless technology, I also imagine everyone will be able to keep a database of their health which they will share with their doctor’s office.

And be sure to check out the video of the new battery in action:

Source: neurogadget.com