Review of Pangborn 2019
This article is part of Aigora's "Review" series, in which we review resources, such as books and conferences, to help our readers find valuable content.
Okay, I'm back in Richmond now following an excellent trip to Scotland for Pangborn 2019. So many things happened I've had trouble putting my mind around it, but here's my attempt at a review. I hope you like it!
Conference Score: 5 stars out of 5
Saturday
I flew into Scotland on Saturday morning before the conference, on a rainy morning. Due to a technical glitch with Verizon (thankfully soon resolved), I didn't have data when I arrived so I wasn't sure where to go. I hadn't even written down my Airbnb location! Fortunately, I ran into Ratapol Teratanavat, who helped me find my way to Edinburgh, and I was able to get my data problem resolved on the tram. Andrew Carnegie once said that capitalism turns luxuries into necessities, and that principle definitely applies to smartphones.
Anyway, I spent the rest of the day getting settled into my apartment and then sightseeing with Will Russ of The Institute for Perception, before having dinner with a great group of friends at the wonderful One 20 Wine Cafe. Pangborn was off to a great start and it hadn't even begun!
Sunday
Sunday morning I allowed myself to sleep off a little jetlag, then headed over to the conference center to register. I found that I wouldn't be able to hang my poster until after the conference had started, so I had a nice meal at MILK restaurant with my friend, Chris Van Dongen. I then headed back to the conference in time for the opening, which was a treat! We were greeted by bagpipe music and then a clear explanation of the logic behind the program by conference organizers Joanne Hort, David Lyon, and Cindy Beeren. In my opinion, this thoughtful organization is what made Pangborn such an excellent conference this year.
After opening, Mikel Cirkus gave an excellent talk on the use of semiotics to spot trends long before they are apparent, somewhat akin to an expert surfer who can read the swell far out at sea to predict where and when high-quality waves will break. What impressed me most about the talk was the idea that a non-trivial amount of valuable information resists digitization, and that one really needs to go out into the world to get access to it. For example, street art can be predictive of long-term trends, but street art data is not easily mined off the internet (at least, not yet).
After this first keynote, there was a break during which I set up my poster. While I was eventually successful in this endeavor, the process took much longer than I expected and I, unfortunately, missed Nathalie Martin's talk on the future of sensory. I heard from several people that the talk was both very good and featured predictive modeling from large datasets among the key tools of the future. The day concluded with the welcome reception, at which the food, networking, and conversation were all excellent.
Sunday night, I had quite a bit of Aigora business to attend to, including filming our weekly recap, so I managed to get that done after midnight and got it sent off to our video editing crew before finally getting to bed around 2a. Here's the video, filmed on location!
Monday
Monday morning I was exhausted from the time change and late-night, so I decided to pace myself and let myself sleep until 8a. While this was certainly the correct choice energetically, it did mean that I missed the morning keynotes. However, I did make it in time for the morning session on consumer behavior, nutrition, and health. For me, the most interesting piece of information was that rats who become obese end up with damaged taste buds. This sort of finding underscores to me just how difficult it is to predict sensory experiences from instrumental and analytic measures. I really think that machine-learning experts who come into our field, without an appreciation for just how contextual sensory data are, greatly underestimate the difficulty of their tasks.
After a nice lunch, I headed off to the MMR workshop, "Free the Spirit," which was an entertaining and well-organized treat. The big-picture message of the workshop was the need for marketing to be connected to the sensory experience that a product actually delivers, a message with which I whole-heartedly agree. In practice, the workshop consisted mainly of historical and technical notes on whiskey distillation by David Thomson, some very pleasant whiskey tasting, and some rather shameless promotion by certain members of the MMR team. In all, though, the workshop was well worth attending.
After the workshop, I presented my poster, "Open Source Tools for Generalizing TURF Analysis," created in collaboration with Will Russ of The Institute for Perception and Frank Rossi of PepsiCo. The poster covered a comprehensive look at how linear programming can be used to revolutionize TURF analysis, allowing us to answer questions that we previously wouldn't have thought we could ask. And, even better, all of this progress can be conducted through freely-available, open-source tools. I enjoyed numerous stimulating discussions and met many interesting people at my poster, reaffirming my belief that a poster is often better than a talk for networking and business generation.
The day concluded with an ASTM happy hour at "The Hanging Bat," a famous and beautiful pub right near the convention center. About 40 people attended and we had a great time, after which I enjoyed Nepalese food with another group of friends.
Tuesday
Tuesday began with a remotely-delivered keynote by Nimesha Ranainghe of the University of Maine. This keynote featured remarkable technology that allowed for simulated flavors to be created via electronic stimulation on the tongue. While the applications demonstrated were somewhat silly, to be honest, the technology is intriguing enough that it will no doubt find applications in emerging categories, such as e-cigarettes, or in fixed-use applications such as drug-delivery where masking bitterness that can't otherwise be eliminated could be important.
The second keynote on Tuesday was an excellent talk by Betina Piqueras-Fiszman of the University of Wageningen on the complex psychological interactions that affect sensory and behavioral outcomes. Once again I was reminded at just how complex sensory data are, with caveats for those that would simply set sensory data as the output of a predictive model.
The third keynote on Tuesday was by Maria Veldhuizen of Yale University, in which she revealed some surprising results involving mismatches between perceived sweetness and actual energy levels of foods. This talk was not the most relevant to my research interests, but it was interesting and well-delivered regardless.
After a refreshment break, I moved on to the workshop on digital opportunities and big data. In this workshop was one of the best presentations of the conference - a presentation by Leah Hamilton, a graduate student at Virginia Tech - on the development of a lexicon for whiskey tasting developed by scraping whiskey reviews online. The research presented was excellent, a fact made all the most exciting by how many ways it could still be strengthened.
Tuesday afternoon brought a welcome break from the conference, with several tours offered to conference attendees. I had signed up for the whiskey distillery tour at Glenkinchie, which afforded me a chance to see the beautiful Scottish countryside. The tour was fun, and I concluded the evening with dinner with friends once again.
Wednesday
Wednesday started on a high note for me, as we had the first-ever AigoraPlus meetup. Six of us met for breakfast at the restaurant MILK just down the street from the convention center, and we had excellent discussion and networking. We even invented a secret AigoraPlus handshake! My only regret about the breakfast is that we didn't take pictures. Next time!
After breakfast, it was back to the conference again, this time to hear Gaston Ares' excellent keynote on sensory methods. What impressed me most about his talk was his emphasis on the need to under your methods well enough to know when to use which method. His presentation was also very clear and energetic, and I left his talk inspired.
The next keynote was one of the best of the entire conference, this time by Ludovic Depoortere. His talk focused on the new technologies that are impacting sensory and that sensory must leverage to stay relevant in the coming era of digital transformation. Since this is exactly my area of interest, I was very excited by the topics he discussed and appreciated the care he put into what was an excellent presentation.
After a break, I attended the Sensometrics session, where the high point for me was the talk that Will Russ gave on our joint research on the use of machine learning to characterize sensory segments. What is most exciting to me about this research is the ability of modern machine learning algorithms to detect complex interactions, taking us past traditional methods such as considering pre-defined consumer groups or looking for differences between sensory segments on a small number of variables of interest. In addition to presenting valuable material, Will did a great job with a clean and clear presentation and was, to my mind at least, a stand out in the session.
In the afternoon, following another nice lunch, I attended the session on technology-led methods. This session was very inspirational as it showcased the use of many new technologies, especially those such as virtual reality (VR) that help provide a more immersive testing environment. I thought all of the talks in the session were thought-provoking in various ways, and I was happy to receive some new ideas.
After the session, I took part in more networking, got ready for the gala dinner, and enjoyed an excellent night including the high point of hearing the famous "Red Hot Chilli Pipers!" (Note, that's pipers, not peppers, get it?)
Thursday
Thursday brought the end of the conference, which was okay with me because I was exhausted by that point. I began the day with a workshop on data-driven modeling taught by Katya Vladislavleva of Datastories. The tool they demonstrated had a very clean user interface, and looked easy to use, yet I still feel that we haven't quite gotten to the bottom of how to leverage the explosion of data and computational power for maximum insight. Following the workshop, I had a helpful discussion with Wim Vaessen of Essensor, and I have some ideas for meta-level modeling that could provide new insights, so stay tuned!
Following a break, it was time for awards - congrats to Hal Macfie and Sara Spinelli for winning their respective awards! It was then time for the final two keynotes - first a talk on so-called "useful data" by Trevor Davis, formerly of IBM, and then a talk on the sensory challenges of feeding astronauts by Grace Douglas of NASA.
I had high hopes for the talk on useful data, as many sensory scientists don't have access to big data beyond publically available datasets that are often highly biased, but the talk fell flat for me as it didn't seem grounded in sensory or consumer science, and didn't contain any big "aha" moments for me. At the end, however, I did appreciate Trevor's comments on the possible ROI of engaging with historical datasets. As for Grace Douglas's talk, it was excellent! While I don't expect to be involved in feeding astronauts any time soon, her presentation was exciting and mentally stimulating. Also, she noted that food might taste different in zero-gravity situations, once again underscoring the sensitivity of sensory data - this sensitivity turned out to be a theme of the conference for me.
The conference then closed, and that was the end of a wonderful week!
That's it for now. If you'd like to receive email updates from Aigora, including weekly video recaps of our blog activity, just click on the button below to join our email list. Thanks for stopping by!