Who: Kevin Bolding

What: Postdoctoral Fellow in Olfaction and Memory

Where: Duke University

Years in the Game: 2

Education: BS in Biochemistry and Genetics

Dr. Kevin Bolding let’s us into his life as a postdoctoral researcher.

Q. What is your insight into being a postdoctoral researcher?

A. I can only compare being a postdoc to being a graduate student. A postdoc is expected to be much more expert at what he or she does and you figure out that you actually did learn a lot in your graduate training about how to solve problems. My opinion seems to carry a bit more weight than when I was a graduate student. Also, you’re totally responsible for whether your experiments succeed or fail, so you just have to keep pushing and make it work- that’s my experience. Maybe that was the case in graduate school too but it wasn’t as obvious to me then.

Q. How would you change your training experience?

A. I wish I had an engineering education because I’m finding that I want to know things about applied math, signal processing, electrical engineering. Right now I have to try to learn these things on the fly. There are a lot of little electronics projects in the lab just to make the experimental equipment work so I’m constantly looking up equations and parts and anything engineering related. I’ve gotten a bit more grounded in signal processing which was easier to adjust to since I played with audio recording a lot before I ever saw an electrophysiological recording. To understand the state of the art in theoretical neural coding and decoding algorithms seems to require some idea about machine learning which rests heavily on linear algebra. I’m having to go back and study just to catch what may seem to other authors as obviously interpretable symbols in an equation.

Q. What is your goal (if postdocs and academics lived in Eden)?

A. My first option was always to stay in academia, find a tenure-track position, and pursue my interests in the neural basis of memory. Academia has a lot of positive aspects in the types of problems you get to engage and the types of people you interact with. Also, the research community is international and gives opportunities to potentially pursue work anywhere in the world, which is enticing to me. However, I’m struggling a little with the idea that academia must consume your whole life. Research funding right now is poor, so competition is fierce, and work-life balance is one of the first niceties out the window. I think the best thing I can do for now is do my job to the best of my abilities which entails continuously developing analytical, programming, and organizational skills which will be useful no matter what I do in the future.

Q. What would you do if you were not a postdoc?

A. My answer to this question changes from week-to-week. I’m motivated by novelty so I would love to be involved in a larger set of shorter projects or projects where I play a less central role. My expertise or my experience rapidly acquiring expertise could be useful in other areas that I am interested in. Immediately adjacent to my current knowledge are medical or consumer applications of electrophysiological data acquisition and analysis. A number of companies are sprouting up and building applications on top of lightweight EEG sensing devices. I don’t have an idea for a killer app in this framework but I can imagine my background would be relevant to someone who does. A step even further away might be to combine my signal processing background with my interest in audio processing in a speech or music analysis context.

Q. What would you like to invent that is music related?

A. I suppose one thing in music analysis that is far from solved is automatic curation/recommendation. I’m sure there are many applications that integrate across your currently library and your social network. I wonder if there is now a space for a more intelligent application that automatically collects feedback about its selections and adjusts accordingly. In public settings, bars or parties, there are many acoustic cues that indicate how a selection was received (tones of voice, singing along, etc.). Music selections could be correlated with public behavior (bar tabs, a force sensor in the dance floor, etc.) I don’t know if anyone is doing such a thing. The same types of intelligent selection might also be useful for the FitBit set trying to come up with the perfect workout mix. I’m sure a lot of people believe that they perform better while they are listening to ‘Eye of the Tiger’, but is that true and if so what do you play next? There would have to be a training period and the algorithm would have to continuously update, but I bet you could improve performance over just picking the Survivor station on Pandora.

Put on some deodorant and check out Olfaction Part I: The Work