Translating earpiece

Foreign languages have always interested me since I started learning Spanish in 5th grade. I think it's so cool when someone can understand and speak multiple languages. It can be difficult and overwhelming though if you have no idea what somebody else is saying to you because you don't speak their language. In the future, however, this may not be as much of an issue. There is currently a campaign in order to raise funds for an earpiece that can recognize different languages and translate them in real time. It's called the Pilot, and it is a fascinating piece of technology. 

The piece works with the combination of speech recognition and typical translation technology. Speech recognition is an extremely complex area of computer science because it involves several other disciplines. Computers start by recording speech and turning it into a spectrogram to make it visual and easier to understand. Each sound shows up on the spectrogram differently. The computer is programmed to recognize these different sounds and piece them together to form whole words. This is a very basic explanation of a very basic kind of speech recognition, but computers can get much more complex. For example, computers can be programmed to predict the next word based on the previous word, in a way, to improve recognition. There are also much more advanced models that work similarly to a human brain to interpret speech. Ultimately, it is an extremely complex process that I can't fully understand. 

a spectrogram of some speech

The recognized speech is then sent to a smartphone app that will create a translation, and send that translation back to the earpiece. The earpiece then dictates the translation so the user can communicate effectively. All of this happens almost instantly so the conversation taking place can maintain a normal flow. 

This technology is so fascinating, although I don't think people will be as motivated to learn other languages when it becomes popularized if they know they can just buy 1 earpiece and app to understand every language.

___________________________________________________________________________

https://www.indiegogo.com/projects/meet-the-pilot-smart-earpiece-language-translator-headphones-travel#/

http://www.explainthatstuff.com/voicerecognition.html

More diabetes management technology?? No way!!

I know I just did a blog post about diabetes management technology, but this new device in the works is much cooler and more advanced. It's called the closed-loop system, or the "bionic pancreas". Before I go more into detail about what it is or how it works, you need some background information about type 1 diabetes management in order to fully understand how cool this device is. Diabetics' pancreases no longer produce insulin or regulate blood glucose levels so we are required to do the work of our pancreases. This includes checking our blood glucose levels via finger prick about 6 times a day, counting how many carbs are in a meal, entering the number into an insulin pump and taking the dosage, or manually calculating the amount of insulin needed and giving it via injection. In addition, if our blood sugar levels are too low we need to eat sugar, or if they are too high, we need to give more insulin. It requires constant thought and management.

Some technology that diabetics can currently use to manage blood glucose levels

That's why the bionic pancreas is so exciting to diabetics everywhere, it does all the thinking and calculating for you! Now onto how it works and how it relates to computer science. The basic closed loop system has three parts: a continuous glucose monitoring system (CGM), a device that can run the complex algorithm, and a special kind of insulin pump. The CGM is attached to the body takes a reading of interstitial fluid. It converts this reading to a measure of blood glucose and sends it to the device that runs the algorithm. The algorithm is extremely complicated and has been in the works for about 10 years now. It computes the proper treatment to give the user based on their blood sugar. Not only does it learn to treat each person's blood sugars differently, but it also takes into account other factors that can affect blood sugar like meals, exercise, illness, hormones, and emotions. The calculations are sent to the pump, and either insulin or glucagon (a hormone that raises blood sugar) are delivered via a small wire in the body. Computer science is crucial to this system because if the dosages are wrong, the results could be extremely serious, even fatal. The algorithm needs to be so precise in order to work properly.

An early version of the bionic pancreas

The most recent prototype of the system. Not pictured: the CGM, a small
patch worn on the skin. The numbers will be sent to the pictured device

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 

(Some more information from personal experience, if you're interested)

Two of my friends who participated in the bionic pancreas camp study!

I'm not going to talk much more about comp sci here, but if you're interested in medicine or this technology, then keep reading! :) I work at a summer camp for type 1 diabetics and for two years (2013 and 2014) they tested the bionic pancreas system! It was so cool to see it in action and how well it works. My friend, Grace, was one of the patients in the trial and her HbA1c (a measure of average blood sugars over time) was perfect which is nearly impossible to achieve with the current technology. The inventor, Ed Damiano, hopes to get the technology approved by 2017 when his diabetic son goes to college. It is looking like it might take a little longer, but approval is on the horizon. After doing all the work for one of my organs for the past 15 years, it would be so amazing to finally have this technology and not need to worry about managing my blood glucose levels. Here's a cool video they made at my summer camp when they first tested it there! https://www.youtube.com/watch?v=06rpdVs0okE

(sorry this blog post was excessively long, I'm just really excited about this)

____________________________________________________________________________

http://sixuntilme.com/wp/2015/11/06/bionic-pancreas-update-call-action/

http://helmsleytrust.org/programs/health-type-1-diabetes/case-studies

https://t1dexchange.org/pages/study-reveals-poor-disease-control-among-adolescents-and-young-adults-with-type-1-diabetes/

https://wefunder.com/betabionics/

IBM Watson can understand natural language

I'm sure almost all of you heard of IBM Watson, the computer system, when it beat the two greatest human champions in Jeopardy history in 2011. Watson is a super computer that can easily answer questions presented in natural language rather than a programming language like java. It was so amazing when Watson won Jeopardy because not only did the computer need to scan immense amounts of information to quickly find the answer to every question, but it also needed to understand and answer the questions in an abnormal way to keep in line with the rules of the game. This was a huge advancement in computer programming, but playing Jeopardy isn't something that happens on a daily basis. Watson, however, has huge real life applications. Before we get more into how Watson is being used, let's talk about how it works for a second.

Watson winning Jeopardy. This picture shows Watson's "thought process" with the confidence ratings

As I mentioned before, Watson can understand natural language, and this is one of the reasons why it is so smart. A typical search engine takes keywords, finds documents containing the keyword, and sorts them in order of relevance, usually based on popularity. Watson, on the other hand, can understand the question in much greater detail to return a more precise answer. Watson instantly scans a 15-terabyte data bank of human knowledge, finds the relevant information, decides on an answer, and when it reaches a level of confidence, presents its answer in natural language. This is pretty much what our brains do, but Watson can store much more information and can complete the process much faster. That is why it is being used in the medical field to treat patients.

One of the most fascinating applications of Watson is in the medical world, particularly in oncology. Watson helps medical professionals make more informed decisions about treatment plans. Using the same process Watson used to win Jeopardy, it scans the analysis the patients' medical records, huge medical databases to find potential evidence-based treatment options, and produces suggestions for treatment with ratings of how confident it is that the treatment would work. Watson has the potential to give better treatments to people while also seeing more clients in a shorter period of time. This is only one of the amazing uses for Watson in our world. I can't wait to see what it can be used for next!


http://www.govtech.com/health/IBMs-Watson-Broadens-Cancer-Crusade.html
https://www.cnet.com/news/ibms-watson-may-provide-a-shortcut-to-treating-cancer/
http://blog.fashionsealhealthcare.com/ibm-watson-impacting-healthcare
https://en.wikipedia.org/wiki/Watson_(computer)#Software

Computer made of DNA?!

Every day, people are working to make the next microprocessor that will be smaller and faster than whatever is on the market at the moment. If these manufacturers keep using the same materials, however, one day they won't be able to make these chips any smaller or faster. Eventually, they will need to switch to making these microprocessors with a new material. Luckily, a new kind of computer is in the works: a DNA computer. The microprocessors will have DNA instead of silicon to store information and complete operations. Although these DNA computers are a fairly new reality, they can already "perform 330 trillion operations per second, more than 100,000 times the speed of the fastest PC." Wow! 

The idea to use DNA in computers was introduced in 1994 when Leonard Adleman thought about how effective DNA is at storing information in our bodies, and how this efficiency can be translated to computers. He proved through experiments that DNA can also be used to solve complex mathematical problems. Because of all of the possible benefits of using DNA in computers, computer chip manufacturers have been working tirelessly to make Adleman's idea a reality. Some of these benefits include:

• unlike silicon chips, DNA biochips will be made cleanly, with no toxic materials
• DNA computers will be much smaller than silicon computers
• DNA computers can hold much more data
• "One pound of DNA has the capacity to store more information than all the electronic computers ever built;­ and the computing power of a teardrop-sized DNA computer, using the DNA logic gates, will be more powerful than the world's most powerful supercomputer."
• silicon computers perform operations linearly (one at a time), whereas DNA computers can perform operations parallel to each other, so complex problems can be solved in hours instead of months or even years!
• an endless supply of DNA, as long as there are cellular organisms on Earth
• DNA is a cheap resource (because of the endless supply)
• 

Although this technology most likely won't be available to common people any time soon (it will likely be used for national governments or huge corporations), it is still amazing to think about!

References:

http://computer.howstuffworks.com/dna-computer.htm

http://news.nationalgeographic.com/news/2003/02/0224_030224_DNAcomputer.html

http://www.calvin.edu/chimes/2014/11/12/dna-computer-chip-technology/

http://images.wisegeek.com/black-microprocessors.jpg

Motion Capture graphics

Motion capture, or mocap for short, is a process that filmmakers and animators have been using to create lifelike animations for movies, tv shows, or video games. It records human movement in real life and converts it to a digital character. It was used extensively in the movie Avatar (2009) to create the Na'vi people.  

As seen in the picture, it involves the human actors wearing suits covered in markers, usually bright lights that are easy for a camera to track and a computer vision algorithm to identify. Cameras will track the movement of the markers, and special computer animation software will identify the location of the markers. Animators will then take this data acquired from the cameras and use it to animate characters that move just like a person would, but that look very different from a human. Computer science is relevant here because mathematical and computational computer programs and algorithms are necessary in order to take just a video of some people in weird suits and convert it to an amazing animation.

This video shows some great examples of what motion capture can be used to animate, it's worth the watch!!  https://vimeo.com/169599296

_____________________________________________________________________

References:

http://66.media.tumblr.com/tumblr_lhbpwwr8PC1qgtvbuo5_1280.jpg (avatar picture)

https://en.wikipedia.org/wiki/Motion_capture#Methods_and_systems (general info about mocap)

http://www.organicmotion.com/motion-capture/ (more info about mocap)

http://www.cs.utah.edu/~halzahaw/MotionCapture_main.html (briefly describes the process of animation using mocap)

Computer program to predict blood sugar spikes and crashes for diabetics

For the past 15 years of my life, I have had Type 1 Diabetes, an auto-immune disease in which your pancreas no longer produces insulin so your body can't regulate it's own blood sugar levels. For diabetics, it is extremely important to maintain normal blood sugar levels to remain healthy (roughly between 80-120mg/dl), even though it is often very difficult to do the job that an organ in your body is supposed to do on it's own. In recent years, however, technology has made managing diabetes much easier. A software program that is currently being developed and tested can predict spikes or crashes in blood sugar, which can lead to seizures, up to an hour before they happen.

The software uses support vector regression to find patterns in blood glucose levels over a period of time and predict future issues. In a similar fashion, support vector regression has been used in the past to predict stock market prices. "Support Vector Regression is a tool from machine learning that can build a regression model on the historical time series data in the purpose of predicting the future trend of the stock price." When graphed, blood sugars over time can look very similar to stock prices over time, so it is easy to see how the same technology can be applied.

Blood sugars from a few days on a graph

From what I can understand, support vector regression works by using previous data and algorithms to generate new data. The system that is currently being developed would use factors such as previous blood glucose patterns as well as user input like "I'm going to go exercise now" or "I just ate a big dinner" and run algorithms to predict future blood glucose levels. After using this new form of software for a while, the computer will begin to learn how each individual person's body and blood sugar levels will react to different stimuli. The more a person uses this software, the more accurate predictions it will provide. This software would be incredibly useful and in some cases even life-saving for diabetics everywhere.

____________________________________________________________________________

References:

http://www.ourdiabetes.com/f/userfiles/image/dawn%20phenomenon.png (blood sugar picture)

http://graphs.net/wp-content/uploads/2009/03/stock-graph.jpg (stock picture)

http://ieeexplore.ieee.org/document/6703098/?denied (quote about SVR)

https://www.ohio.edu/research/communications/diabetes.cfm (article about blood sugar software)