Robotics, Earthquakes, cancer research….Intel?

Intel Research Pittsburgh Lab opened doors today to academia, press and just about anyone in the general Pittsburgh public who wanted to stop by. The open house buzzed with enthusiasm from an estimated nearly 200 attendees about the advancements the lab is making and ideas sparked on potential areas for collaboration.

The lab is one in a network of university labs Intel opened back in 2002 along with Intel Research Seattle and Intel Research Berkley. There are roughly 22 full time Intel employees at Pittsburgh, plus a revolving door of Carnegie Mellon University students and professors that are jointly working on projects together. This collaborative research model has a primary intent of publication, that is, to open source and share findings.

The research looks out approximately 5-10 years and focuses on projects like personal robotics, healthcare technology research, earthquake simulation and many others. The research sounds far-out, and you might even wonder, what does all this have to do with Intel? The answer is actually not so subtle. Each project is looking at creating real world applications of the future that are going to put more and more demand on computing. Today the applications are running on clusters or in a simulated environment, but they will all take advantage of multi-core, tera-scale and parallel computing, crunching tera-bytes and even peta-bytes of data. Each project proves that this industry is still quite young, and consumers have so much more to realize and benefit from technology and their interaction with technology. The findings provide insights internally to Intel business units, and they are shared with universities, and typically are published in leading forums.

There were more than 70 projects highlighted in the open house today that show how an increase in computing power and capability will enable new ways of living with computers. The researchers have put together some very compelling demos. I’ve highlighted 3 for you below, and will be sure to return to the blog with more in the near future.

The Robotic Barkeeper: “Personal Robotics” is a joint project between CMU, Intel Research Pittsburgh, Intel Research Seattle and the University of Washington. The ultimate goal is to make robots more human-like vs. those in a factory, able to assist and deal within uncertain and unstructured environments like the home. The video below is showing the demo of the Robot “waiter” (or robotic segway) going around the room, avoiding obstructions on it’s own and allows you to set cups on the tray. The waiter delivers to the “barkeeper” (or robotic arm) so it can grasp the cups and place each one appropriately in the dishwasher. This is a significant step in personal robotics because in every demo variables change, and the robot figures it out on it’s own. (there’s no human behind it instructing each move) Watch the video to see it in action:

Earthquake simulation:

One of the emerging “themes” that Intel Research Pittsburgh will focus on is this idea of “big data”. There is so much data available to the world that are stored electronically but are unused. Google is a great example of a company putting data, in their case internet and satellite data, to use. But it’s just the beginning. In this example, the researchers are developing a ground model generator using Map/Reduce, a parallel programming model promoted by Google. earthquake.jpgTheir goal is to enable seismologists to more accurately, quickly and cost effectively evaluate the impact an earthquake of different magnitudes will have on a given region, California in this case. They can determine what will happen to the bridges, buildings, and oceans in a matter of hours instead of weeks. This demo is running on 50 servers with dual quad core processors each, and is working through peta-bytes of data (that’s 1 MILLION gigabytes!) This is the first example demonstration as this Big Data concept is just beginning….


Skin Cancer Diagnosis:

One of the healthcare related projects is using technology to assist healthcare providers with skin cancer diagnosis. The team is currently working with the University of Pittsburgh Medical Center (UPMC) in a trial experiment. Thinking back to the big data concept – there are many hundreds of thousands of skin cancer cases with medical records and images of blemishes documented and sitting unused. Technology can put these images to use to help doctors make more informed quick decisions about a patient. A camera tool is used by a doctor to capture the image of a skin blemish or mole. It is then entered into the software tool and searches through a database of other images, recognizing those that are most similar. Note, it’s not looking for just titles or descriptions – but actually looking and analyzing the image, just like a human would. This takes huge amount of computing – again crunching terabytes of data in very short period of time. The application finds those images in the database that are most statistically similar and can allow the doctor to compare in detail specific examples w/the current patient situation. The doctor now is empowered to make an even more informed, data driven decision with percentage examples of what might happen with their current patient. This project is run on system called Diamond (downloads here) which provides parallelized implementation of search tasks.

There are so many more projects to highlight – this is just a sampling. According to some of the CMU professors, this type of open source industry relationship is incredibly rare. In fact, they couldn’t think of another joint industry lab of its kind, so open with the findings and results. This model is a win-win for everyone involved. Intel is able to extend research resources beyond the 22 employees to the many brilliant minds of students and professors who are eager to participate through internships and personal time and interest. Carnegie Mellon students and faculty are provided access to resources (like a $200,000 robotic arm), and grounding in what really can be done in the industry. Working with Intel gives academics the opportunity to potentially see their research translated into real-world technologies. Being on the edge of Carnegie Mellon University Campus is critical, the lab is thought of to many as another department of the university.

Comments are closed.