There is no conception in man's mind
which hath not at first, totally or in parts,
been begotten upon by the organs of sense.
---Thomas Hobbes, Leviathan, 1651

Thursday, June 2, 2011

Graph Theory, Flavor Combinations and Menu Development

What is a food combination?  It could be flavor combinations as in flavored yogurts, or also as in ice cream toppings or add-ins, or toppings on a pizza.  I've had an interest in scientifically modelling such combination since I was at the CIA.  Flavor combinations are as much an art as a science.  Chefs spend years learning, developing and refining these combinations - this process is never finished.

What models do we have?  Well, there are methods to evaluate pairs of flavors.  For example, you can easily give people two flavors and ask them to give you a number from 1 to 10 on how well they go together.  But that is boring, and not predictive.

An early study that excited me about flavor combinations was the development of a novel scale by Margaret Cliff and Marjorie King.  In 2005[1] they investigated the quality of wine and cheese pairings.  Their scale starts with a 10cm horizontal line.  On the far left side are the words "cheese dominates."  On the far right side is the phrase "wine dominates."  In the middle is the phrase "ideal pairing."  A subject would taste the wine and cheese, and then mark a line on the scale indicating their perception.  By assigning numbers (after the fact) to intervals along this scale, quantitative comparisons can be made to evaluate the quality of the pairings.  Using a bit fancier statistics, you might even able to be predict the qualities of ideal wine or cheese to pair together to hit that "ideal point."

Though this method is great, how do you properly evaluate more than two combinations?  Well, extending the King and Cliff methodology, you could employ an isometric grid (e.g. triangular grid) to evaluate the quality of three items (wine, cheese and cracker for example), but I suspect the people doing the evaluations would go cross-eyed.

Initially I tried to use the Napping method to answer this problem.  However, the methodology was not the right model.  So I put the thought on the back-burner and continued my research in multivariate statistics.

Figure 1: Combinations of salad toppings
Sometime in the summer of 2009 I had an idea.  Using graph theory, the math behind social network analysis (e.g. Facebook) I could model food combinations.  Graph theory uses two simple constructs: nodes and edges.  By modeling different flavors, or foods, or menu items as "nodes" and what these items are connected to in a compatible combination as "edges" the problem of finding a suitable model for food combinations is solved.  In a collaboration with John and Danny Ennis at the Institute for Perception, we put together a research program to investigate the suitability of this graph theoretic approach.

Figure 1 on the left shows a visualization of a graph related to a single person's preference for salad toppings.  Here are 14 toppings, represented as nodes (the black circles) and their pairwise compatibility represented as lines.  For this individual, blue cheese and broccoli would go well on a salad, but blue cheese and corn would not.  (To get this information, we simply asked the subject about the binary (yes/no) compatibility of all possible pairwise combinations of these ingredients).  The magic occurs when we use this visualization to predict larger combinations.  For example, the trio of carrots, onions and tomatoes all have connections between them, so the prediction that this person would like a salad with these three toppings holds true (which we tested, its true).  The same holds true for predictions of four, five and six topping salads, as long as all possible connections are present.  The beauty of this method is that we don't actually have to ask people about all possible combinations of three, four, five and six toppings to get this information (that would be almost 6,500 potential salads for 14 possible toppings), yet we get information about all of them!

The power and potential of this methodology was so exciting to me that I switched my dissertation topic a year and a half before graduation.

Perhaps the greatest contribution made with this method was in regard to the Army Field Rations known as MREs.  MREs contain 11 different food components such as entree, side, snack, bread, spread, dessert, snack, etc.  The Combat Feeding Directorate has many available items within each of those 11 categories.  So many, in fact, there's over 22 billion potential MREs.  By employing this graph theoretic approach outlined above we surveyed soldiers and applied the techniques to MRE menu development.  In the end, we presented the Combat Feeding group with the best optimal MRE menus given the available combinations.  This project was the capstone of my dissertation.

Happy graduation!

[1] King, M., and Cliff, M. 2005. Evaluation of ideal wine and cheese pairs using a deviation-from-ideal scale with food and wine experts. Journal of Food Quality, 28(3), 245-256.

Thursday, May 26, 2011

How to Cook Pasta

Just a quick note to point out that I was featured today on the incredibly useful and popular blog "Lifehacker."

Click here for the story.  Lifehacker is where I usually search first when I want to know the optimal way to do anything with electronics, but they've recently branched into food and I hope to continue helping them with their articles.

Thanks Jason Chen!

Sunday, May 22, 2011

NSRDEC, CIA, SSC, Ph.D. and other important acronyms

Almost a year to the day this unexpected hiatus is now over.  The past year has been one of the busiest yet, and as those of you that been to grad school (or have significant others that went) you probably understand that it is not a 9-5, especially that last year.

In short, since we spoke last I've conducted research for, and defended my Ph.D. dissertation (I'm walking next Sunday back in Ithaca!), moved to the metro-Boston area, planned and begun execution of a series of multivariate statistics workshops and started work a couple weeks ago at the U.S. Army Natick Soldier Research, Development and Engineering Center (NSRDEC) as a Postdoc working alongside the Consumer Science folks.  Just a couple of weeks ago, I was asked (and accepted) to join the Culinary Institute of America's Research Advisory Council with the likes of Harold McGee and Paul Rozin - I've mentioned numerous times that the CIA holds a special place in my heart and I'm ecstatic to be able to support their research mission in this manner.

The Natick Soldier Systems Center (SSC), of which NSRDEC is a part of, is one of 13 or so RDECs scattered around the US.  Natick is the only RDEC that focuses on the soldier's interaction with the world.  Clothing and textiles, ergonomics, food and shelter are all major research areas (among others).  New gun?  Its no good if it causes blisters after a couple minutes of use.  What about a new Heads Up Display?  Its no good if the soldier is overwhelmed trying to process thousands of pieces of information, or if it weighs to much, or if the batteries don't last.  New gadget?  How do you balance the weight with the hundreds of other gadgets we'd like our soldiers to carry.  This is where we come in.

The U.S. Army has a long history in both food and consumer science research reaching back to the 60s when the Quartermaster Institute was in Chicago.  The Combat Feeding Directorate (which I am not a part of, but do work with) is charged with developing rations for the U.S. Army. (If you'd like to jump ahead in time to a future post topic, go here). The Major accomplishments are numerous and stem from the fact that the Army allows its scientists to conduct both basic and applied research.  The former components of that is huge in the research world - as sensory scientists, there's few opportunities anymore to conduct any sort of basic work.

Besides supporting other functions within the Consumer and Cognitive Science teams, I'm quite excited about my personal research track which will focus on food and emotion (something completely new and different for me).  If you follow the sensory literature, emotion is a very big deal right now, yet all of the work is observational thus far and focuses on how to measure it.  We're working on the why and how aspects of food and emotion.

Welcome back everyone!