A Robot Passed the Self-Awareness Test and This Is How It Did It

When talking about robots and self-awareness, I think most people would just freak out, but there are some people who would be extremely excited and interested about these things. But I don’t think freaking out would be the case here, even though a robot just passed the first self-awareness test ever.

The guys over at Rensselaer Polytechnic Institute in New York are said to have built three robots which were put to the so-called “wise men puzzle” test. The original test involved a fictional king, who in order to choose his next advisor, invites three of the land’s wisest men to a contest. He then puts either a blue or white hat on their head and tells them that the first to stand up and tell the colour of their own hat will become his new advisor.

The same logic has been implemented with the three robots here too. Two of them were stripped of their ability to talk, then all three were asked to tell which one was still able to speak. All of them then proceeded to say ‘I don’t know’, but surprisingly, the one who heard its voice became aware of its ability to still speak and added ‘Sorry, I know now!’.

The above may seem trivial to us as humans, but bear in mind that robots are programmed to do what we ‘tell’ them to do, so up until now, all robots we’ve seen doing this were doing them because we systematically told them what to do. To see a robot recognising its own voice and distinguishing it from other human voices is a big step forward for AI.

To be noted here is that all three robots were coded in the same way, so we can see a bit of the machine learning technique in place here. While the other two did not see any signs of self-awareness and ‘thinking’, the third one was able to tell and learn the difference. This means that the third robot was able to learn some differences in behaviour, using the base code to ‘deduce’ what the others could not.

It’s really interesting and I admit, it may be a bit scary too. But in the end, complex AI are bound to be invented sooner or later, so we may see the first big step here. Also, if you’re interested in seeing the robot for yourself, you should know that it will be displayed at the RO-MAN conference in Japan between the 31st of August and the 4th of September.

Thank you TechRadar for providing us with this information

We Could Be Teaching Computers How and What to Learn in the Near Future

We’ve seen a lot of projects and methods involving machine learning up until now, but the truth is that no matter how efficient and quick a computer is, the software will always be its limit. However, Microsoft plans to change that and is working on a new tool that would allow you to teach computers new things.

There is a lot of information about the project and tool on Microsoft’s website, but let’s get straight to what the Microsoft Research team wants to do. Take a jigsaw puzzle for example, where every piece of it fits somewhere and contributes to the big picture once complete. This seems to be the approach Microsoft hopes people will take, where each jigsaw puzzle piece is created by someone knowledgeable in his or her field. This means that a doctor can teach a computer how to search for specific patient details in a database, while a chef can teach it how to make the best dish according to a recipe.

However, to get a lot of people from different fields together in this project, Microsoft also needs to build the tool as simple and understandable as possible. This means that Microsoft is looking for a way to make the tool more autonomous and offer a simple and understandable user interface. I mean developers know how to use such tools, but they can’t expect a doctor to know how to use a complex development tool.

The tool, named Language Understanding Intelligent Service or LUIS for short, is part of the Oxford Project. Though it still is in its early stages, it may prove to be a strong and useful tool for complex AI development. You can view more details about the project over at Microsoft’s Blog. The tool can also be found over at its website here, but you will need to get an invitation first to try it out.

Image courtesy of Radio1.be

This Is Why New Software Comes With Old Flaws

You are probably wondering why we hear that legacy flaws are still present in new software. Well, the answer is simple. Developers have a habit of reusing old code for most of their projects and the code is not reviewed for all potential flaws, but rather the approach tends to be similar to the slang ‘if it works, then don’t try to fix it’.

This does not mean that developers are lazy. The approach is favourable even by top-notch programmers because of the tight deadlines they have to meet, so time will always be above everything else when shipping new software.

However, this comes at a hefty price. While we hear of many hacking incidents, only a few of them are complex enough to break even the most impenetrable systems. Most of them were done by exploiting the already ‘implanted’ flaws in all software products. Everything except the operating systems can be deemed ‘hackable’ by most people with some knowledge of hacking.

The flaws go so deep that even some government departments are at high risk. Security analyst found out that some software in government departments is still based on older programming languages. But is this the future of programming? Of course not.

Security analysts in the field say that the problems with legacy flaws may likely increase, but they don’t have to. The real problem is that, by focusing exclusively on shoving new software on the market, companies forget about security completely. A better approach here is to split project development into two major components, development and testing, which could work in parallel. This way, a lot of bugs could be fixed and major security bugs flagged before the software hits the market.

Thank you CNET for providing us with this information

Image courtesy of nikopik

Science’s Biggest Problems Solved By … Gamers?

It is unbelievable, but the truth is that the plain old gamer can really make a scientific breakthrough. Scientists might ponder, think through logically, mathematically and even metaphysically about humanity’s unsolved problems, but it sometimes takes a little craziness and another point of view, of course, to get on track and find the key element in solving a problem. And this is where gamers come in.

It is reported that in 2011, people playing a game called Foldit, an online puzzle genre game about protein folding, solved the structure of an enzyme that caused an Aids-like disease in monkeys. Researchers began working on a solution 13 years before the game launched. The gamers, however, solved it in just about three weeks.

A little over a year, another group of gamers playing Planet Huners, a space exploration game, found a peculiar looking planet with four stars in its solar system. And since then, around 40 planets that could support life were found, all of which have been previously overlooked by astronomers.

When you first hear ‘gamer’ and ‘scientist’, you think that they are far apart. But when you look closely, both parties are looking for the same thing, either in a lab or at home on a screen. Both gamers and scientists are looking for the same thing, a pattern that matches, whether in logical puzzle games such as Candy Crush for example where you have to find the matching candy shape and color to destroy blocks and collect points, or in a scientific environment where you have to find a pattern in a DNA string that causes mutations or diseases.

“Our brains are geared up to recognise patterns,” says Erinma Ochu, a neuroscientist and Wellcome Trust Engagement Fellow at the University of Manchester, explaining why scientists are turning to gamers for help, “and we do it better than computers. This is a new way of working for scientists, but as long as they learn how to trust games developers to do what they do best – make great games – then they can have thousands of people from all around the world working on their data.”

However, don’t get this wrong, but not every gamer can become a scientist. Logical games are aimed to get your brain up and thinking about a solution to a basic problem explained in a graphical environment, a game. Gamers are known to have a keen eye for detail and come up with the simplest solution to a complex problem, hence the ‘new point of view’ that scientists want. Whereas a scientist who is being cooked up in a lab somewhere with complex problems and calculations on his mind tends to miss that simple solution.

Thank you The Guardian for providing us with this information