Deep Learning Neural Networks

Every single time you decide to update your status, you are sending secret messages about your personality that you yourself may not be able to realize. I’m not writing about just the basic facebook like of your favorite band’s page. Companies like facebook are now analyzing the data you harmlessly give them a bit further to understand where you are in your life and what is best to sell you tomorrow. The larger companies are investing large amounts of money into something called a deep learning neural network with the hope of figuring out just who you are and how to best proceed knowing their customer’s personality.

How does this deep learning neural network figure out who I am and what I want? The idea is based on the concept of how our bodies figure things out and make decisions based upon the inputs we receive. The input values can be as simple as your facebook post that you really want some ice cream and your later post that you had a really hard test this morning. These post can be perceived just the same as you seeing ice cream and your decision to purchase it. With the use of the neural network, the computer can realize that during finals time it should to try and sell you ice cream. This is just a very simple example and the technology itself can be used to look at thousands of post to make even more surprising statements about you. Basically, neural networks can be used to recognize associations between what you do with your social network to determine how best to serve and advertise to you.

A_simple_neural_network_with_two_input_units_and_one_output_unit

The associations work nearly the same as neurons in your own central nervous system. Each neuron is attached to thousands of other different neurons. To determine outputs from inputs, your brain takes your neurons and their associations to other neurons to make complex decisions. Computers can do the same thing, but replace the word neuron with node. Nodes can store information themselves and also connections to other nodes. Deep learning algorithms take these nodes into consideration and evaluate the connections between the data to provide meaningful results.

Although all of the data mining of your personal life may sound scary, neural networks do have serious potential to make your overall social networking experience much better. They hope to utilize deep learning neural network algorithms for a variety of applications such as helping you better articulate a post and choose the best vacation pictures from the thousands you have uploaded.

The idea itself isn’t brand new. Computer neural networking technology was used by Google to recognize human faces and helped Microsoft with speech recognition and translation. The same idea has even been used to recognize human emotions from text without you telling it specifically how you are feeling. One company who was later acquired by facebook had claimed they could accurately determine which of your smiles were genuine and which were faked just by looking at your pictures. It is a very efficient and intelligent method of analyzing data and making valuable connections in any application of its use.

November 3, 2013 at 11:55 pm Leave a comment

Touch Technology

Touch technology is pretty advance now a days.  The sensors that they use on phones and tablets are just incredible.  I know on my phone (Samsung Galaxy S4), while I’m watching a YouTube video, if I look away, it will pause for me until I look back.  It is one of the eye recognition sensors that the Samsung has developed.  This has spread from smartphones to tablets and also laptops as well.

One thing that these touchscreen devices are lacking is feedback.  You cannot feel the buttons you are pressing when you move candy’s over in Candy Crush.  Nor can you feel the buttons when you are typing a text message on your phone.

Well look no further folks! The research department of Walk Disney Company are developing something called “haptic” devices.  Haptic technology is a feedback technology that used the sense of touch by applying forces and vibrations to the user also known as tactical feedback.  This allows users to take control of virtual objects.  If this definition does not help you or if you want to know more, look at quick link below that will explain to you what it does.

I’m sure majority of you guys have played with the Xbox Kinect.  Imagine how much better it would be if you could feel the items around you.  The product Disney is developing shoots a puff of air at you in conjunction to what is happening in the game. As you saw on the YouTube video.  The TV had soccer balls coming at the guy and the sensor device was shooting little puffs of air at him so that he can feel it.  This device that Disney has made is called “Aireal”.

Many other research company’s are developing similar devices to this one.  Be on the lookout in the next decade for something to come out similar to this.  The feature of technology just keeps getting more advance as we know it.  I still think some people are still trying to get on to the whole touchscreen part of it and just wait until this new “Haptic Technology” stuff gets out.  Lets just hope it is as user friendly so that people aren’t left behind in this world of technology.

Happy Halloween and have a safe weekend!  Maybe if this video gets enough likes and comments, I’ll post a picture of my funny Halloween costume! haha

Don’t forget to take the quiz! A couple questions from there might be from the YouTube video. 😉

October 31, 2013 at 11:38 am Leave a comment

SCADA Stories to Tell in the Dark

Alright, campers, gather ‘round. Have you ever heard of the acronym SCADA? It stands for “Supervisory Control and Data Acquisition”. SCADA is pretty important stuff; it encompasses all the systems that control our cities. Our electrical grids, our traffic control lights, our water supply and sanitation systems, even our power plants.

So now — the sun has set, there’s a chill in the air and it’s almost Halloween. Huddle up to the warm glow of your computer monitors and I’ll tell you all a ghost story…

In fact, you may have already heard this one. It’s the one about a hacker who created a sophisticated self-replicating computer program that could spread itself — completely undetected — into the controller hardware for nuclear facilities around the world. It was beyond stealthy, even deleting itself if necessary. It would spread virulently, but become totally inert on systems that didn’t meet its exacting requirements, waiting for further opportunity to infect other hosts. Once inside a desired nuclear reactor host computer, it could dangerously modify the operation of the reactor equipment while at the same time sending bogus statistics back to the operator’s terminal, telling him everything was fine, until it was too late…

The scariest part, though? This actually happened. The worm I’m referring to was known as Stuxnet. There have actually been several versions of the Stuxnet worm, but the one I described was used to attack Iranian nuclear power plants. Stuxnet never caused any meltdowns, not that it couldn’t have with the level of control it was able to achieve, but because it was programmed not to.

But why make Stuxnet in the first place? No one can say for sure, since no one knows specifically who made the software, though it’s likely that the development of the software was funded by a government; the software is so sophisticated and requires such a deep understanding of the targeted systems, it would have required an incredible amount of time and effort to create. What’s worse, analysis of the spread of the worm seems to indicate that the initial release and spread of that strain of Stuxnet was unintentional. A programming error caused that version of Stuxnet to spread beyond the targeted plant, and go on to infect nearly sixty percent of all the computers in Iran. The worm has even been found on a small percentage of computers in the US. To top it all off, there have even been reports of software very similar to Stuxnet being sold on the black market.

Here’s a video that goes into more detail about Stuxnet if you’re interested. It’s really a fascinating story.

Situations like Stuxnet are quickly revealing how truly insecure so much of our technology and infrastructure is. Stuxnet targeted Iranian nuclear power plants, but those are far from the only vulnerable SCADA systems in operation today. Stuxnet was likely developed by a highly-skilled team over the course of several years, but as software and technology progresses, the skill barrier that once stood in the way of a task like Stuxnet will begin to evaporate. Even now, you can use a simple search engine to locate potentially vulnerable SCADA interfaces on the internet.

capture

A quick, dumb search reveals dozens of potential targets.

All of this means that, in the coming years, there is going to be a steady increase in the demand for security professionals to design and implement more secure systems; so if that seems like something you would be interested in doing as a career, you might want to look into Network Security as a major… before it’s too late!

October 28, 2013 at 12:05 am Leave a comment

Chinese instruction manuals

Okay, so you’ve just bought some gadget made in China, and you can’t figure out how to use it, or put it together, without taking a quick look at the instructions. But your gadget doesn’t carry a big brand name, like Apple. Instead, it’s made by a small exporter who doesn’t want to pay a native English speaker to write the instructions. Trying to read the instructions leads either to frustration or laughter–whoever wrote them was clearly not proficient in English. Probably most of us have had this experience, although I think the situation may be getting better over time.

Machine translation from one language into another is motivated by problems like these. People sometimes want to be able to read a Web site not written in a language they understand. Or they want to do business with people who speak only another language. Or maybe they would like to travel to a country whose language they don’t know.

Computer scientists and linguists have been working for years on trying to get computers to translate from one language into another. And they have made tremendous progress from the early days. You can get apps for mobile devices that do a good job of translating short sentences not involving novel word usage. Some browsers have built-in translation ability for Web sites, and these sometimes produce results that are good enough to make sense of the site.

But language processing is one of the most challenging aspects of artificial intelligence. When we write things to our friends, we assume all sorts of unspoken things, like nicknames we use for others, or a context that determines which meaning an ambiguous word has. The machine doesn’t know about those things, and will probably screw up any translation of such things because of that. Suppose you say, “I’m going down to the bank with Big Fuzzy to try to catch something.”  You are probably not going to a financial institution to make a withdrawal, and only your friend knows who or what Big Fuzzy is.

Google Translate is one of several free translation services available on the Web. Try it out right now to see how well it works for you, at translate.google.com. Just type some English sentences into left box on the site, and choose a language for the right side box, and click “Translate”. The translation appears in the right side box. Maybe try Chinese. Then, you probably want to check to see how well the system worked. You can just click the “Switch” button, to make the translation go in the other direction, back to English, but I don’t trust that method as a good check. Instead, to try to prevent the system from “cheating”, copy the Chinese translation into the box on the left, and run the translation back into English on the right. See if you get back the English that you put in at the beginning. (Chinese is a good choice if you want to make the system work hard, since the grammatical structure of Chinese is very different from English.)

If your English sentences weren’t too complex, the system may very well work well for you, and that would be pretty impressive, assuming you know that word-for-word translation simply does not work. The machine has to do a lot more “thinking” than just looking up individual words in a big dictionary.

To give Google something harder to work on, I asked Translate to put

“All living creatures classify objects and aspects of the world, in the sense that they treat some stimuli as more alike than others”

into Chinese, and then translate the Chinese back into English. This is what I got:

“All taxonomic objects and aspects of the world, in this sense, they treat some of the stimulus are more similar than others”

So, maybe that instruction manual for your gadget wasn’t written by someone who understands English poorly. It might have been Google Translate that wrote it. Apparently there is still some job opportunity for human translators!  And there certainly is a job opportunity for human beings who love language and who understand something about how computers work, to help make systems like Google Translate work better. And that would be a good thing, I think.

If you are able to generate a funny translation of something using Google Translate, why not share it in the comment section below?

Kenton Machina

October 24, 2013 at 2:16 am Leave a comment

Balloon Powered Internet Access

Have you ever needed to check your email or Facebook while camping, hiking, or any other area where there was no chance you were going to get access to the internet? Google aims to fix this. The same lab that is working on the Google glasses and the Google car are also working on a very different project. They want to provide WiFi globally with the use of balloons.

In most areas of this country, it isn’t too difficult to find a WiFi connection. In other countries it can be a different story. Internet access can cost more than a workers monthly income. Imagine holding a second job just so you can shop amazon. In other places, there is absolutely no availability to get on the web. Google claims that for every one person that can use the internet, there are two that can not. Project Loon intends to provide an inexpensive way to grant access to these places, as well as every other place on the planet. Imagine what internet lacking areas can achieve with just this simple connection. Children can get an education where they couldn’t before. Farmers can achieve greater harvest with better access to weather data . You could even stay connected to your professional life while on a camping trip with your family in the mountains.

The main idea is to send these balloons into the stratosphere where Google can steer and control them with the use of wind current and solar energy. The balloons will sail around 20 km in the sky (commercial airplanes fly around 10 km). The stratosphere’s wind normally flows from west to east, however there is ability to use different wind directions to provide some level of control. This allows the Google team to steer only by lifting and dropping a balloon in the different levels of the stratosphere. Keeping the balloon stationary would cost far more than sending many balloons operating in unison over a network, so Google will send multiple to circle the planet in a ring formation. When one balloon leaves an area, another is meant to take its place providing access everywhere constantly and inexpensively.

The technology is still in its very early stages and much work is left to be done before this becomes a reality. At its current state, the speeds are estimated to be similar to our current 3G speeds. While it isn’t perfect, it will provide a solution to a complex problem that we face in this technological era. If we can provide constant communication to everyone on the planet, imagine what else could be achieved by looking for answers in places we never have before.

October 20, 2013 at 11:55 pm Leave a comment

Cloud Storage

You may have heard people using terms like the cloud, cloud computing, or cloud storage. But what exactly is the cloud.  It is pretty much a hard drive that you can access anywhere you go without having to bring your personal computer.  When something is in the “cloud”, that means it is stored on servers on the internet, instead of on your computer. It lets you access your documents, music, videos and more from any computer that has an internet connection.  There are big warehouses filled with servers where all of your data is stored.

There are many reasons to use the cloud, but a couple of the main reasons are convenience and reliability.  At this moment, most people especially students like you carry a flash drive around with data that is important to you.  But how many times have you has to replace a flash drive because it was lost or suddenly stopped working.  I know I have been through at least 5-10 different flash drives because of those reasons.  So no more forgetting your flash drive plugged into public computers or in your pocket where you will find in the washer later. Saving a file to the cloud makes sure that you’ll be able access it with any computer that has an internet connection. Also, when you want to share a file with a friend or for a group, it is easy to give them access to the files you want them to see.

With the cloud, you’re much less likely to lose your data, since it is stored on servers. If you have a bad track record where your laptop/desktop usually dies and you lose all your data, the cloud may be for you.  Another thought is how secure is it? As long as you set up a good password, you should be fine.  Is there really a big difference to losing your flash drive versus getting your account hacked into? I’m sure there a lot of advantages and disadvantages to both but in the end they are pretty much up to par.

MAT 120 students! Don’t forget to take the extra credit quiz and stay dry out there!

Here is a video to explain the cloud more in case you still do not understand.

http://www.youtube.com/watch?v=RNWHgWg8iGg

Have a good rest of your week!

David Nguyen

October 17, 2013 at 9:44 am 2 comments

The Thing About the Internet of Things and Other Things

It’s been about a dozen years since I passed the fourth grade. That probably doesn’t make me sound particularly old — as it shouldn’t, that’s not the intention of this anecdote. You see, what boggles my mind is that, around that time, it was still often considered unfair for a teacher to ask that a paper be typed on a computer, as a lot of families still didn’t have computers, let alone printers or even — hold the phone (literally) — the internet. Compare that to today, where practically every person you cross on the street is likely concealing one (or several) considerably powerful computers that are perpetually, wirelessly connected to the internet.

It seems a bit old hat to go on about ‘how far we’ve come’ and things like that, though. We’ve all heard it at least a dozen times, computers are getting more powerful and ubiquitous every day — so what else is new?

Well, despite an amalgam of issues with infrastructure, data throughput, storage and dwindling radio spectrum, the next evolution of the ‘Internet’ is still rapidly approaching. There are a number of different interpretations on what that evolution may look like, but all of these theories describe what is generally referred to as an ‘Internet of Things’.

If you’ve ever heard people joke about how ‘pretty soon, everything is going to have an internet connection’, then you have more or less been introduced to the concept of an ‘Internet of Things’. Basically, an ‘Internet of Things’ is an idea of a society where practically all ‘things’ that are bought or sold by humans will be connected to the Internet in one form or another.

Now, to me, it seems that there is a bit of a stigma associated with this concept. People see it as unnecessary, excessive, or in some cases even a little unnerving in a ‘Big Brother’ sort of way.

bigbrother

I wouldn’t say people are wrong for thinking that way — it’s an important thing to consider as we continue to push the boundaries of technology — but I just wanted to play a little devil’s advocate here and talk about a few of the potential applications of this technology that have the ability to do a great amount of good.

For instance, during an emergency situation like a building fire, first responders could be equipped with software to scan for victim’s phones or other internet connected devices in order to ascertain their location. In addition, all of the networked objects in the bulding could be used to allow the firemen to determine the safest path to the located victims, or perhaps to provide them with data on where to concentrate fire suppression for maximum effectiveness.

A more pedestrian, everyday use would be something along the lines of a ‘smart house’. Your fridge would read the RFID tags attached to all of your food products and automatically compile statistics like expiration dates and generate recipes based on the contained food. Your phone would signal to your house when you arrived, whereupon it would turn on lights and set the climate.

In fact, Corning Glass’ series of concept videos for their idea of a ‘smart home’ are a good example of what an ‘Internet of Things’ could look like within a home network.

Most of the technology needed to enable these sorts of interactions with ‘Internets of Things’ exists today, the only thing needed to bring it to life is a proper implementation (easier said than done, of course, but that’s not to say you can’t start messing around with the tech yourself). If anything, I hope this might help you look past the potential downfalls of an ‘Internet of Things’ and see the kinds of incredible advancements in software and technology by incorporating these concepts into our everyday life.

October 14, 2013 at 12:05 am Leave a comment

Older Posts Newer Posts


July 2017
S M T W T F S
« Dec    
 1
2345678
9101112131415
16171819202122
23242526272829
3031