Although we keep pushing forward, I thought this would be a great time to reflect on 2011 as well. I’ve done a ton of blogs around my technical work with the Big Data team and our events, so I wanted to focus on our group of Edison Engineers here at GE Global Research. It’s been an exciting 2011 within the program, and we’ve jump started 2012 with the hiring of four new Edison members already!

Last year we had approximately 25 Edison members at any given time, with 18 in our Traditional track, and 7 in our Software track. We graduated two members, Andrea (fellow EdisonsDesk blogger) and Matt (author of the infamous “light saber” blog entry), who have taken full time positions here at the Research Center. Congratulations to both of them!

As a group, the Edison members have been extremely active in writing, producing 46 publications both internally to GE and externally through conferences and journals. We have also helped to drive the future of not only GE, but of the industry with 19 intended and 15 filed patents  having an Edison member as either the primary, or co-inventor!

Of course, as the saying goes, we work hard and play hard – the Edison team has had plenty of fun adventures this past year. Usually, we all try to get together once a month, holding pot luck dinners, heading out to movies or just getting together to watch some football. At the end of the 2011, we all took an afternoon to do some charity work for the Habitat for Humanity. This was a great opportunity to get our hands dirty, helping the community while learning more about ourselves.

As we look towards 2012, I can only imagine bigger and better things! We could see a record number of graduations, with the potential for 5-10 graduates this year (myself included) and hope to expand our Software track as well. I hope you all had a successful 2011 as well and are looking forward to an even better 2012! As always, feel free to reach out to me through the comments below, on Twitter (@EdisonsDesk, @CTMcConnell) and remember to check out what our GE Experts are working on at www.ge.com/Experts.

Until next time,
Chris

I wanted to put up a quick blog to mention something that I think will be pretty exciting throughout 2012.  For those of you who have been following my blog you know that I’ve always included my Twitter handle at the end of my posts for you to be able to connect with me on.  However, recently, many more of my colleagues at Global Research and at GE have begun to pop up on Twitter.  I wanted to call out all of my fellow Global Research Tweeters in this blog, be sure to say hello and check them out!

Adam Rasheed:@AdamRGE
Sumit Bose:@Sumit_Niskayuna
Kristen Brosnan:@KristenBrosnan
Charles Theurer: @CharlesTheurer
Tim Richter: @Tim79mc
Andrew Reid: @AndrewAReid
Ambarish Kulkarni: @AmbarishJK
Carlos Haertel: @CarlosHaertel
Steve Gustafson: @StevenGustafson
Juan De Bedout: @JuanMDeBedout
John Lizzi: @John_Lizzi

Also, don’t forget that you can find me at @CTMcConnell and you can follow the blog at @EdisonsDesk! You can find a full list of “GE Experts” on Twitter by visiting www.ge.com/experts.

The International Conference on Case-Based Reasoning (ICCBR ’11) was held in Greenwich, London September 12-15, 2011. I attended the conference along with Bill Cheetham from Global Research and Deniz Terry from GE Intelligent Platforms. This conference was a great opportunity since both my projects currently involve case-based reasoning.

The ICCBR had four different sessions: The main session consisted of a single track of talks and posters. The second session included a variety of parallel workshops. The third session was an industry day where companies were invited to talk about their applications of CBR. The fourth session consisted of two parts 1) a computer cooking competition where people design algorithms to make the computer come up with interesting recipes given some ingredients and 2) a doctorial consortium where ten Ph.D. students presented their thesis topics and received feedback from senior researchers in the CBR community.

Some of my favorite talks from the conference were the ones that presented new open areas of research in CBR. These were: “Assembling Latent Cases from the Web: A Challenge Problem for Cognitive CBR”, “Case-Based Learning by Observation”, and “Amalgam-Based Reuse for Multiagent Case Based Reasoning”. One of the major open issues that these papers presented is that CBR currently requires a lot of effort in creating a case base and maintaining it. Some of the proposed solutions to this would be to make the CBR systems identify missing cases and learn by observing an expert. Another solution is to make the CBR extract the information from the web. This also brings up another new research area in CBR which is to learn from multiple sources and also learn multiple tasks. Of course, allowing the CBR systems to learn on their own presents whole bunch of issues such as determining whether the sources of information is reliable, how to assemble cases from multiple sources, how to limit the amount of information learned, and of course reducing the amount of time it takes to search for information and assemble the cases. So there are a lot exciting things going on in the CBR community.

During the industry track Bill Cheetham talked about 12 different applications of CBR at GE. Many people were quite impressed about how much CBR is used at GE and the amount of money CBR has saved GE. Another interesting talk was about IBM Watson and the application of CBR in structure mapping of “Jeopardy!” clues. And it is interesting to see that many of these big projects such as IBM Watson are using technologies similar to ones that are already developed and in use at GE!

Members of GE Global Research won a bid to host the 2013 ICCBR Industry day and doctorial consortium at GE Global Research. The main session and workshops will take place in Saratoga, NY.

These webinars will provide you, our future world changers, an overview of GE Global Research, the Edison Engineering Development Program and then go into some specific details around projects that our own EEDP members have been working on! I can tell what you’re thinking, “There are two webinars, which one is right for me?” Let me shed some light on the two Webinars and even a little detail on what the EEDP is.

First, let me describe the EEDP in case you have not yet heard of it. The EEDP is a GE wide leadership program, which includes technical training, professional training and multiple (3-4) rotations in different areas of the business. Now, when I say leadership, a lot of people think “manager,” however the EEDP is designed to create new technical leaders. Senor and Principle technologist that help define, shape and execute/implement new advances in their area of expertise! There are some differences between the Global Research EEDP and our “businesses” (i.e., GE Energy, GE Aviation, etc.), so let me highlight those quickly:

• For our program, applicants should have at least a Master’s degree. We have hired PhD level candidates in the past, however we are slowly moving away from that (sorry Doctors!).
• At the Research Center, your rotations are typically centrally located. This is because the Research Center works to bring all of GE together, so under one roof you can work with GE Healthcare, Transportation, Power & Water and more!
• The technical and professional training is different than that at the Businesses, but I won’t spoil the surprise, as our webinars will talk about some of this.

If you want some more information about the EEDP in general, feel free to visit here, or throw some questions my way.

Now that you are all excited about the EEDP, which webinar should you attend? Well, our traditional EEDP typically hires the following backgrounds: Aeronautical Engineer, Biomedical Engineer, Chemical Engineer, Computer Engineer, Electrical Engineer, Materials Science, Mechanical Engineer and Physics degrees. Anyone with a degree in those areas would be best suited to attend the first webinar (again: November 8^th, 3PM EST).

Students that are studying Computer Engineering, Computer Science, Electrical Engineering and Mathematics might be better suited for our Software track, (November 15^th, 3PM EST). Now – this was a test; did you see what I did there? I’ve repeated Electrical and Computer Engineering because depending on your area of research/interest, you might be viable for either track. For example, we have an EE now that has a background in Machine Learning and Artificial Intelligence. We also have EEs that work in Embedded Systems design and Controls. If you’re unsure, I’d encourage you to attend both webinars.

Of course, if you are unable to attend the “specific” webinar for your background, but can make the other one, please feel free to do so! The webinar is being run by current EEDP members and they would be more than happy to answer your questions – some bring some good ones! Once again, you can register for the webinars here. Feel free to ask questions and leave feedback about the webinars below as well, and of course, you can always reach out to us on Twitter as well (@EdisonsDesk, @CTMcConnell).

Until next time,

Chris

For those of you that haven’t heard of the software webinars (shame on you! Just kidding ) some of our lead scientists, lab managers and project leaders took some time over the summer to discuss the interesting projects they have been working on. We’ve discussed topics such as Supercomputing, the “Internet of Things” and Social Media and Collaboration. The webinars were so successful that we’ve decided to bring them back for another round!

The previous webinars and registration for the upcoming webinars can be found at www.ge.com/softwarewebinars so make sure you head over there to catch up on our previous sessions and see what’s coming soon! The first installment in our new series will be entitled “Movement Planner and Trip Optimizer” discussing the work that the Research Center has done with GE Transportation to analyze and optimize freight train patterns on rail networks. This webinar is scheduled for October 11^th at 3:00p.m. EST., but don’t worry, we’ll record it and post it on the site if you can’t make it for the live date.

The Research Center has a really cool lab associated with this work where we’ve actually built a simulation system like an arcade game where users can act as a train conductor to speed up and slow down the locomotive engine on a given route. The simulator measures the fuel utilized, the trip time and “wear and tear” on the locomotive which is compared to our algorithms’ suggested usage to see how a human operator compares to our optimized plan. Its pretty fun, but I can speak from experience; it’s frustrating that the algorithm always seems to win!

Anyways, make sure you register for the webinar, (again, October 11^th, 3:00p.m EST.) by clicking here! As always, the success of these webinars starts with you guys, so bring your challenging questions, imaginative thoughts and enthusiasm as you join the presentation. If you have any thoughts on topics you’d like to see us talk about, comments about past or upcoming webinars or general questions about Software at GE Global Research, post comments below or find us on Twitter (@EdisonsDesk, @CTMcConnell).

Until next time,

Chris

You can register for the upcoming webinar below, or visit the main site here. While you are there, check out the other webinars: “The Internet of Things” and “Supercomputing for Engineering and Scientific Discovery,” which have been recorded and posted for your viewing pleasure!

As always, we are looking for feedback on the webinars and the Blog itself! Let me know what you think either through comments below, email or Twitter (@EdisonsDesk, @CTMcConnell).

Until next time,
Chris

The forum, which takes place August 4^th and 5^th, is the second annual forum held here are Global Research. On the first day, we invite representatives from many of the GE Businesses as well as vendors and speakers from external companies. We then take the second day to provide the GE Business representatives insight into what our researchers have been doing on the BigData projects; detailing our thoughts on where technologies are headed and what we believe the next big thing will be – spoiler alert, that’s where I come in. The event is wrapped up with a Q/A session and general discussion among all GE attendees to discuss the forum, current and future problems and what the next steps for the company could entail.

As a part of the BigData team I’ve been working heads down with my colleagues to finish up our proof of concept utilizing in-memory and on disk solutions for handling BigData. We’ve also been busy testing, benchmarking and researching all sorts of new technologies to provide the Businesses detailed information about how the technologies work, what they are good for and how they can be utilized by our business teams to solve the problems of today while expanding to the problems of tomorrow.

Our two hour (yes two full hours…whew!) presentation will include some amazing demonstrations utilizing real-time analysis and highly interactive machine learning and analysis. These actions are then combined with historical lookups across billions of entries within seconds. The best part, the analysis is done through light-weight systems allowing users to utilize a desktop, laptop or mobile device (tablet/phone).

While our team is a piece of the puzzle, the external speakers and vendors always provide a new, interesting perspective on the industry trends and demands. The list of vendors is long, but it includes (in alphabetical order) Amazon, Cloudera, DataMeer, EMC, Flumebase, Hadapt, IBM, Informatica, Karmasphere, Microsoft, Oracle, Progress, Rainstor, Revolution, SAP, SAS, Splunk, Syncsoft, Talend and  Teradata. We also have external speakers (again, in alphabetical order) from American Electric Power, Cycle Computing, EMC, Health and Analytic Innovations, Medtronic, Progress Software, O’Reilly, SAP, and Third Nature.

After the forum, I’ll give you guys a recap of the talks, some up-coming technologies and anything else that really stood out! Follow me on Twitter as we lead up to the forum for some live tweets about the day one activities in-between talks and running around helping with the fun! As always, if you have questions, thoughts or comments, please reach out via the comments section, Twitter (@EdisonsDesk, @CTMcConnell) or email!

Until next week,
Chris

Homework. If there is a single feature of A-Course that sticks out in the mind it is the homework. There are thirty in total, each unique, and requiring a matching write-up. You mark off the time not by the number of the classes passed but the homework assignments left.

Like all events, some assignments make a bigger impression than others. Similar topics and related work may blur together but there is almost always that one that will stand out. For me, it was the second half of the Decision Science homework.

The scenario was fictitious: fake failures for hypothetical product lots, set up in realistic distributions. Our task was to take this data and find a cause for the failures, make recommendations on a course of action, and wrap the whole thing up as though it was a report to senior management.”

Each week, one site had to present their solution to the rest of the software students and the assigning instructors. This week was our turn.

If you have ever taken a statistics course, you are familiar with the basic structure of such assignments. They are the no-bake cookies of mathematics. The student is given ideal data, just to show that they are aware that analysis can be done and which method to use. Ten minutes later, one is done with a sense of satisfaction. This was certainly not the case for the Decision Science assignment.

The data was a mess. We were given four data sets covering the captured failures. The data was incomplete. Units were missing and mismatched. We were not even certain how many product models there actually were. Information about impurities, shifts, bore size, and the car models themselves were mixed and matched, preventing a consistent view of all the failures. Far from ideal, this data felt real.

Remember the ten minutes that it would ideally take to a solve a similar problem with ideal data? This took 30 hours or so. Working with my partner, Aisha, we ran through the Minitab litany. The results were not encouraging. We transformed the data, playing with the units involved, converting between metric and obsolete measurements, and generally trying to improve our results. After five hours, we could not find a collection of variables that seemed closely related to the failures. Statistically significant variables weren’t practically significant and vice versa.

The next 25 hours were a dream of repeated analysis and novel approaches. I’d list them, but reading them would get you as close to the answer as trying got us. For clarity, that is 25 hours, 6 hours at a time, evenings after work.  Frustrating? Yes.  Interesting? Absolutely!

The night before we were set to present, we still had no definitive answer to possible causations so Aisha and I improvised. We went with that as our answer.  The assignment was phrased that we should report out as if we were presenting to senior management.  Neither of us would have gone to an actual manager and suggested making changes based on such results, so we presented our lack of results as a result. We gave our best candidate causes and then advised not throwing time or money at pursuing them. We turned it into a process capability issue. Instead of talking about fixing the problem, we advised a course of action that could get better data and find an actual root cause. In the end, everything worked out.

At the time, I really liked this assignment. The uncertainty of the presentation appealed to me. I got a lot of practice at basic statistical methods chasing ghosts and dead ends but I think the really appealing part was that, where we could have just been given a series of short exercises or one massive one that took forever, we were given something that was typical of real life. It was challenging and engaging without seeming sadistic or abstract to pointlessness.

Probably least important, our presented solution was accepted as valid and a good way to deal with the situation.

Great news everyone! This week we’ll be holding our next software webinar on Supercomputing for Engineering and Scientific Discovery! I’m personally very excited about this topic, for a few reasons. First, in a previous blog post I described the Advanced Computing Lab, where I did my first rotation in the Edison Engineering Development Program. This lab was managed by this week’s presenter – Rick Arthur.  Second, the idea of High Performance and Scientific computing is ever growing, with the use of GP/GPUs, commodity clusters and high end super computers. Finally the work that is done using these machines and techniques have enabled GE to make decisions in a fraction of the time that it once took which I have to say is really cool.

If you haven’t registered yet, don’t panic! You can click on the image below any time before Rick starts the webinar, which will take place July 12^th, at 3:00pm EST. If you can’t make it, bookmark the main site to revisit and watch the recorded version later!  You’ll find out why images such as the blue and green one above, could change the future of computing.

As I mentioned in my last plug of “The Internet of Things,” if there are topics that you would be interested in, post some comments below or reach out to me (@CTMcConnell) or @EdisonsDesk via Twitter. We’d love to know what you all think about the webinar series and what changes you’d like to see!

Hey everyone – I wanted to take a moment to plug our upcoming Software webinars that GE Global Research will be holding over the next 2 months. If you haven’t heard about this yet, take a look at our site for some of the details. I’m really excited about this series, as it will give you guys a chance to hear about some hot topics within the Research Center – plus meet some of our researchers! Make sure you bookmark the site, because we’ll be recording the webinars so you can re-watch or watch it at your own leisure. Before each webinar, I’ll post a small blog about the webinar, the host and provide some technical details about what will be covered!

Our first webinar will be hosted by Joe Salvo and is titled “The Internet of Things.”  Joe is the manager of our Business Integration Technologies lab and is acting manager of the Intelligent Networks lab.  Joe selected this topic as “The Internet of Things” promises to create networks of unprecedented scale and complexity. Join Joe Salvo to learn how society and industry may be transformed by this new system-level innovation.

I hope you all can make it to the webinars! If there are topics that you would be interested in, post some comments below or reach out to me (@CTMcConnell) or @EdisonsDesk via Twitter. Let me know what you think about the webinars as well!

Until next time,
Chris

As you may (or may not) know, the Edison Engineering Development Program involves multiple rotations, three or four, in the range of 6-12 months each. My previous entry about what a “typical” day is like was from my first rotation, however I have recently made a change into my second rotation, a Big Data based project! Although I can’t divulge too many specifics, I thought I’d give you guys an example of what I work on related to this project every day. Okay, maybe not every day, but most days at the very least!

As you can imagine, GE has a lot of equipment in plants, laboratories, hospitals and in the air (just to name a few of course ), all of which generate data. This data might be highly structured, similar to what you’d find in your typical Relational Database Management System (RDBMS) to completely unstructured text, images and video. Not only do we need to find the best way to store this data, we also want to find ways to learn from it and do so in a fast, reliable and scalable manner! Oh, by the way, did I mention that we are talking about data which is in the petabyte range?! It is easy to see that this data deluge can become a daunting task.

That’s where our team comes in. I am working with a team that has approximately 10 people directly related to the project and upwards of 30 that are indirectly associated. For the primary members, it’s our job to research the latest and greatest in Big Data management techniques, whether that’s a traditional SQL approach, the “NoSQL” type approach, or even technologies that fall under the umbrella of “NewSQL.” Okay, maybe we are getting too technical, but for those Computer enthusiasts out there, use your favorite search engine to look up some articles comparing those techniques. For those of you that are interested, I’ll get you started; here is the first link I obtained in a quick Google search. This is a short article that provides a rather complete listing of technologies within the space and I can tell you that on any given day, I get to play with a lot of these technologies!

As I mentioned before, we are looking at ways to obtain results in a fast, efficient manner, something that is counter intuitive when we discuss storing large amounts of data on disk (ignoring Solid State Drives of course). In an effort to increase speed and productivity, our team is also looking at potential techniques for handling the data completely in memory and pushing “old” data to disk only when it is no longer necessary. For those of you that have a background in machine learning, you can imagine the countless problems trying to determine what data is old, when it should be considered old and even if it is old, how we can store it so we can re-access it later for historical comparisons. Our team is hoping to put all of these problems to rest, no pun intended.

Did I mention that most of the data (anywhere from 1-10 Terabytes a day) will be sent in real time through data streams? A portion of the Big Data team is also investigating Complex Event Processing (CEP) tools that can help sustain the high data bandwidth, while providing real-time analytics to track the incoming data for anomalies, errors and even provide prediction if a failure is about to occur. I think a great example of this is within Smart Grid. Think about a city where every home is sending out meter information every minute, 30 seconds perhaps even at the second level. As this data arrives our system(s) need to analyze that data and determine if more/less power should be routed to the city, perhaps even issue warnings to the utility companies that there is a problem and a blackout could occur.

These are the problems that I’m working on every day here at the Global Research. With all of these challenges, you can imagine my day is pretty diverse. Much of my time is spent reading on new technologies, new algorithms published at the latest conference and best practices / techniques for utilizing these Big Data solutions. I’m also in contact with multiple vendors (sorry, no spoilers in this article friends) in an effort to bring their technology to the Research Center and perform benchmarking and design a proof of concept around our different use cases.

It’s amazing how fast a day around global Research goes, but it’s always a blast to come in and start looking at the next emerging technology. Combine all that with the fun activities within the EEDP (like Whitewater rafting recently), everyday just brings new challenges and excitement!

Until next time friends,
Chris

To begin the journey, we travel way back in time (okay, about 2 years ago) while I was still in graduate school. My professor for High Performance Computing (I and II) is a researcher here at Global Research in the Advanced Computing Lab and he really made Global Research sound like an interesting place to work. His stories about GE and some of the public projects he had worked on really intrigued me and his insight into the Research Center made me feel like this was an opportunity to launch a research career, not just a job.

With this in mind, I began searching www.ge.com/careers for entry level positions and came across the Edison Engineering Development Program for Software Developers. After some researching of the EEDP (www.ge.com/eedp/) position, it became clear to me that this position would provide the breadth of knowledge I was looking for, while fine tuning my research skills. After submitting my resume, I played the waiting game…until another EEDP member (and now my recruiting buddy!) Rebecca gave me a call.

As an EEDP member, we are trained by Human Resources to help grow the program. This includes resume reviews, values interviewing and, for those that are super motivated, recruiting trips to find the next class of EEDP members. For those that feel uncomfortable assisting in recruiting you would not be forced into helping – although it is encouraged!

After a few rounds of phone interviews, I was brought on site for a full day of interviewing. The day was actually extra stressful for me. The typical interview would go from 8:30am-3pm, however I needed to be back to school before 3pm for presentation on my Masters’ work, so my interview was condensed by about an hour by removing some of the usual breaks. I left feeling confident but was still (very) nervous to hear the response. Finally, a few days later, I received word that they were interested in hiring me for the EEDP and now, I get to share my story with the world!

As a member of the EEDP, my daily routine is as predictable as the lotto…okay maybe not that drastic; however no day is the same and I love that about the Research Center. When I first started, I worked with GE Aviation and Lockheed Martin for my first rotation. We were looking into next generation computation in aircraft systems – a topic that I knew nothing about before starting at the Research Center. One of the greatest aspects to the EEDP is that when you do your rotations, you’ve got the chance to go into an unfamiliar area and learn something completely new.

Within the first year as an EEDP member at GE Global Research, you’ll undergo technical training (the A-course) as well as some professional training. All members of GE Global Research must go through Six Sigma training within the first year of employment, which is about a week long and teaches everyone how to properly define and measure success of a project.

One final and most important aspect to the EEDP that I want to mention is the fun, external benefits to the EEDP. Just about every week during the summer, some, if not all of the EEDP members will get together for outdoor cookouts, hikes, game nights, dinner or just relaxing in the sun. During the winter, we have had some bowling events, pot luck dinners and have organized trips to see plays as well! Last summer, the entire EEDP went on an Offsite Retreat to a ropes course, which had zip lines, rock walls and all sorts of obstacles to overcome! Coming up soon, I’ll be wrapping up the A-Course and to celebrate, about 15 of us will be going on a white water rafting trip!

I hope that this blog gives everyone a feel for the first year of the EEDP, as well as a look into some of the fun aspects of the position! Even though the position is demanding and requires some excellent time management, the benefits are significant! As always, if there are any questions, please feel free to comment here, or reach out to us on Twitter (@EdisonsDesk, @CTMcConnell)! Keep checking Edison’s Desk for new blogs about some of the GE Software labs and (hopefully soon) a video diary to highlight another day for me here at the Research Center.

Cheers,

Chris

This entry will detail our Computational Intelligence Lab. I recently made a rotation change into this lab plus another EEDP member (Aisha) has been a part of the lab for just about 9 months now. As in my previous post about the Advanced Computing Lab, I hope to shed some light on the fun, innovative software research that is done at Global Research to the extent that I can disclose! If you’d like to chat further, you can add comments below, or catch us on Twitter (@EdisonsDesk, @CTMcConnell). Let’s jump right into it!

The Computational Intelligence Lab  is composed of members with a variety of backgrounds including Computer Science, Computer/Systems Engineering, Electrical Engineering and Mathematics, just to name a few. Some specific technological areas of interest include machine learning, text and natural language processing (a hot spot!), semantic technologies and massive-scale data processing (“Big Data”).

The lab has its hand in a plethora of projects right now (and they are looking for more amazing Edison members, so get those resumes in!) which include:

1.       Big data – I’ll talk in a little more detail about this project because both Aisha and myself are involved

2.       Mining text reports to identify trends, anomalies or patterns

3.       Cyber-security

4.       Social and information networks

5.       Medical claims fraud detection

As I mentioned above, both Aisha and myself are working under the umbrella of Big Data. The project is…well, awesome to say the very least! Aisha is looking at the Big Data challenge from an Artificial Intelligence perspective.  She is working on algorithms that deal with dimensionality reduction, distributed processing and pattern detection techniques.  Her work will help GE’s customers to better understand anomalies, find new patterns within current data and find better ways of handling the processing of that data.

I myself am involved with…well everything of course! Just kidding, my role is to investigate some potential technologies that can solve GE’s Big Data needs.  As you can imagine, the problem becomes significantly harder because of the size of GE, let alone the size of our data. Some of the work I’m dealing with includes machine learning techniques, platform exploration/training/benchmarking and data management strategies.

I hope that this post has provided a little more information about another one of our awesome software research labs! Some of the basic software technologies that are utilized throughout the lab include Java (both applications and web based), SQL and “NoSQL” based technologies, Map/Reduce programming and some popular text mining software/frameworks such as Lucene. The lab’s projects are always exciting and provide Edison members the chance to make a large impact not only within the Global Research, but with the businesses that rely on the Computational Intelligence Lab for new technologies.

Until next time,
Chris

While I hope to continue the “Day in the life…” series, I also wanted to talk about some of the labs that our current EEDP software members are involved with and what types of (super) cool software projects GE is working on. Of course, I can’t disclose all of the information, but if you have questions, feel free to reach out! Add some comments below, or on Twitter (@EdisonsDesk, @CTMcConnell) and we can chat further!

The first lab I wanted to highlight is the Advanced Computing Lab –which was the lab where I did my first rotation! Before I begin, I suggest you watch this YouTube video, just for kicks!

This was a “just for fun” project that some of the lab members worked on in their spare time, but it highlights some of the skills and interests that the group has…plus it’s pretty funny.

In general, the Advanced Computing Lab looks at …well advanced computing techniques. Although the focus is not on hardware, a deep understanding of how computer architectures work is essential for getting the algorithms and applications to meet performance specs. For example, CUDA and OpenCL are programming languages that are used to create programs which run on non-conventional processors like Graphical Processing Units (GPUs). A GPU can be used to achieve highly parallel, extremely fast mathematical processing by taking advantage of hundreds of processing cores. To achieve such fast processing the software/algorithms team need to understand how these cores are laid out on the GPU, how these cores share memory and how to efficiently transfer data between the CPU and GPU.

Other areas of interest within the lab include Message Passing Interface (MPI) programming, data intensive simulation, algorithm optimization, exploring disruptive and emerging architectures and hardware systems and evaluation of these systems. By utilizing combinations of these technical areas (and others), the ACL has been working on a number of really exciting projects such as:

1. Radiation detection
2. Next generation computation on board aircraft
3. Finger print analysis
4. Large scale data intensive simulations of alloys, turbine designs, etc.
5. Biomedical imaging

This lab has been great to work in, the members of the lab love what they do, which makes it easy to learn new technologies as an EEDP member.

In the next lab spotlight, I plan to provide you all with a look at the Computational Intelligence Lab (CIL). As a quick tease, the CIL is involved with projects spanning many areas including Analytics, Machine Learning and Network sciences to name a few.

Until next time,
Chris

November 30, 2010:

As with most days, I strolled into the office a few minutes before 8am so I could get my favorite parking spot and have a few minutes to review my emails before the day began. I knew today was going to be a busy one, but couldn’t wait to get it started! The day began with a visit from my Ph.D. advisor who arrived around 8:30am. Some of the work that he and I have an interest in is related to BigData problems, similar to those that Global Research is attempting to tackle as well. After showing my colleague around the facilities, we had a meeting at 9am with some senior members of the BigData project here. This was a meeting that I had brought together so that both sides (Global Research and my advisor) could discuss some of the common problems they were facing and think about potential solutions!

After the meeting was completed, it was time to meet with a few of my EEDP colleagues. We were scheduled to interview candidates for EEDP today and wanted to make sure we had an appropriate set of questions. As an EEDP member, you go through a set of training by Human Resources called ‘Hiring the Right People’ which enables current EEDP members to interview candidates for the next class.  This is an exciting part of being an EEDP because with this training you will help to shape the future of the EEDP and its members.

All of a sudden, the morning was over and it was time for some lunch. A quick stop to the onsite cafeteria for a “create your own salad” and I was energized and ready for round two! Once I returned to my desk and a few emails later, it was time to do some work on my project. That day I was focused on some work with GE Aviation Systems, working on a report out that I was going to give at the Michigan site on Thursday. Our project team was leaving Wednesday morning, so any last minute touch up work needed to be completed for the presentation. After a review of the presentation, some updates and clean-up it was getting close to 3pm and time for the hiring wrap-up!

During the wrap-up all interviewers get back together to go over the details of the day and discuss the all of the candidates. All of a sudden, it was 5pm, and I wondered where the day had gone. Knowing the next few days would be filled with travel and presentations, I decided to head home to finalize the A-course homework for the week; Systems Architecture. As I walked out to my car a little tired, a little wired, I couldn’t help but think how lucky I was to have this position where the possibilities are endless.

Clearly, not all days are as busy and diverse as this particular date was, however I wanted to give you all a sense of the variety and opportunities that are available! As of this writing, I have been in the EEDP for 5 months, yet I was establishing meetings with external sources and project managers, was preparing for a presentation to customers in Michigan and helping to shape the future of GE and the EEDP class members. The best part about that day: you will have the opportunity to do all of this too!

As Thomas Edison once said, ‘Genius is one percent inspiration, ninety-nine percent perspiration.’

Until next time,
Chris

I am currently in my first rotation in the Advanced Computing Lab, working on projects for GE Aviation Systems and Lockheed Martin. We are currently investigating some potential hardware solutions to meet the new demands in the airline and defense industry. Thus far the projects have been amazing and provided great opportunities to make an impact early. In the beginning of December, I went out to Michigan to meet with the customers at GE Aviation Systems to provide my analysis from the last few months and pitch new areas of research for next year and beyond!

As of this writing, we are almost half way done with our A-Course geared towards Computational Research and Software Engineering. The classes have been really interesting, but the best part of the classes is the networking! An awesome advantage with the “Software A-Course” is that each class is taught by a GE employee at different sites! This provides amazing insight into how the other businesses utilize concepts from Design and Risk Analysis, to Algorithms and Design Patterns to meet their demands, plus we get to (electronically) meet GE engineers that are experts in the course topic.

So what’s next? I’ve got a few more months in the ACL to wrap up some work with GE Aviation Systems and Lockheed Martin before I transition to my next rotation. Right now, it sees as though I’ll be doing some work within our Big Data projects, perhaps working on Smart Grid problems as well. The second half of A-Course is coming up fast with some great topics; Designing for Parallel Processes, User Interface Design and Emerging Software Technologies to name a few.

I hope this post provides an access point for those of you interested in Software and Computational research at Global Research! Keep checking for my blog posts, as I hope to start a running set of entries detailing a “typical” day at work here. Along with that, I hope to provide some insight into the labs that many of our future Computational EEDP members could find rotations in.

Until next time,
Chris