UVM brought a TEDx conference (Independently Organized Technology Entertainment & Design Event) to the Davis Auditorium on 10/28/11, focused on complex systems and entitled “Big Data, Big Stories”.   The 11 presenters together provided an impressive breadth of content, with 5 external researchers from institutions including MIT & Cornell, as well as 6 researchers/scientists/mathematicians/professors from UVM.  You can find a quick snapshot of their bios here: http://www.uvm.edu/~tedxuvm/?Page=speakers2011.php.  These researchers are delving into some impossibly complex questions about humans, our planet and the universe we live in.  Here is a short sampling of some of the huge ideas the TEDx conference covered in one short afternoon:

• Peter Dodds – set the table for the conference, helping the attendees frame up the definition of complex systems as evaluation macroscopic behavior that has no central control.  Peter noted that with the computational power that we have built as a society, we can now (1) store massive amounts of data, and (2) simulate and model an impressive series of scenarios to derive patterns and takeaways from that data.  Peter then offered up the image of a hotdog as an analogy to Collateralized Debt Obligations (“CDOs”), noting that the individualized parts of each of these entities you probably wouldn’t want separately, but assemble them together in a package and then you all of a sudden have an appetite for about 2 million of these beasts.
• Chris Danforth – has developed a model to turn off the effects of the Industrial Revolution, which shows that the 1 degree Celsius temperature increase we’ve seen from 1900 – 2000 disappears from our environment when you remove that human influence from the virtual equation
• Austin Troy – has noticed that we are drowning in imagery & remote sensing data from sources like Google Earth.   Meanwhile, humans are great at recognizing patterns, his example was flashing up a picture of the Big Lebowski next to Eddie Vedder, allowing everyone to recognize how quickly humans are able to tell these two people apart, but how difficult it might be for a computer to do so.  Therefore, Austin has been developing huge rule sets for interpretation of data to train computers how to understand images and think spatially instead of pixel by pixel, so that computers can, say, recognize the difference between a rooftop and a road when surveying Google maps images. 
• Mike Schmidt – is working on automated knowledge discovery, also known as a “robotic scientist”.  He has developed a model named EUREQA that you can download here to detect “equations and hidden mathematical relationships in your data” and distill the mathematical fundamentals of that data.  Mike’s model then helps humans ask the right questions around huge unruly data sets.
• Gary Johnson – is focused on Ecosystem Services, which, for example, places a value on the non-material elements of a forest.  Not the maple syrup or lumber products that come out of that forest, but the oxygen, the recreational value and the nutrients retained in the soil that help prevent erosion and keeps sediment out of the streams and rivers, etc.  Using an Ecosystems Services value database Gary is working to better understand who receives ecosystem services from where, and can that information help us with land management decisions, allow us to evaluate various development scenarios, and interpret the highest and best uses of the land around us.
• Hugh Garavan – is investigating addiction and willpower in the brain during adolescence, looking at the ages at which teenagers take their first alcoholic drink and what that tells us about their future behavior.  Hugh asks the question, “How do millions of neurons work to produce coherent behavior, and what does that tell us about impulse control?”
• Josh Bongard – recently received a Presidential Early Career Award for Scientists and Engineers (PECASE) from Barack Obama.   Josh has been working to regrow the Tree of Life on UVM’s supercomputer, then simulating other potential Trees that might evolve on distant planets.  
• Rob Axtell - His most recent research designs a macro-economy from millions of interacting agents, with a view towards using modeling to help inform policy decisions that might help us recover from future recessions faster
• Marta Gonzalez– is using mobile phones as data sensors to map humans’ most frequent travel patterns, aggregating data points that could assist with traffic flow and urban planning.  
• Isabel Kloumann – believes that the evaluation of the happiness of the human race has been overshadowed by more easily measured metrics like GDP growth and Income Inequality. She is looking at the data from 50MM Twitter users, and scored 10,000 words to evaluate happiness over time through social media.
• Neil Johnson – believes that humans “do” the big events like the stock market crash, that is, humans are the force behind the black swans.  He has been modeling the Foreign Exchange (FX) market which breaks apart and reforms correlations and relationships to other currencies on a constant basis.  Using Agent simulation, Neil believes that he has developed a model that incorporates the statistical properties of currency prices which can predict price crashes in the future.

I came away from the TEDx event thinking that UVM is at the forefront of some pretty powerful and innovative complex systems research.  The work going on at the Gund Institute and in the Computer Science department, including the work with UVM’s supercomputing capacity and the Spatial Analysis Lab is a key asset to Vermont.  Hopefully that’s indicative of the great things that will continue to emerge from our state’s only research university.