Archive for November, 2010

h1

Antarctica

November 20, 2010

I had read parts of a book by Rudolf A. Raff, The Shape of Life: Genes, Development, and the Evolution of Animal Form (1996), over the summer.  One passage particularly struck me: “As a young man, Apsley Cherry-Garrard spent two and one-half years as an unpaid assistant zoologist on Robert Falcon Scott’s fatal 1911 expedition to Antarctica and the South Pole, then returned to England in time to see the start of World War I.  He survived military service in Flanders, and found the time to publish in 1922 one of the most extraordinary accounts of polar exploration ever written, The Worst Journey in the World.  There he vividly and movingly tells of tragic fate of Scott and his polar party, but within that grander tragedy he records the staggering difficulties of one of the strangest trips ever undertaken, his own appalling winter journey with ornithologist Edward A. Wilson and Royal Navy lieutenant H. R. Bowers to Cape Crozier to collect the eggs of the emperor penguin.  The same zeal for scientific discovery that lured Darwin and so many other great scientific voyagers drew Cherry-Garrard and the other members of the party on a journey of scientific exploration that they barely survived.”

Some people say that performing experiments in wet labs is “something every kid that has ever dreamed of being a ‘scientist’ wants to do.”  But for me, undertaking a quest for discovery always held a place in my dreams.  Being the first to see the unseen and to feel the unfelt, all achieved through an arduous journey, would have been the ultimate achievement.  Indeed this human experience is arguably the reason to undertake scientific expeditions.  Within the debate currently raging about space exploration, David Mindell and his colleagues have argued that “human spaceflight achieves its goals and appeals to the broadest number of people when it represents an expansion of human experience.”

So now to give the disappointing resolution to my buildup.  I have no particular plans to go to Antarctica in the near-term.  I cannot justify it on scientific grounds, so the National Science Foundation or its sister organizations around the world will not send me.  I think I will have to wait, save up some time/money, and go on a pleasure cruise.

Now why is it, you may ask, that I have no scientific justification.  Couldn’t a mathematical theory of something be better developed in Antarctica than in Westchester County?  But the whole entire epistemic premise of mathematical theories is that they be universally applicable.  The mathematical theory of communication is designed to be true whether you are in Gaylord, Michigan, or in the Amundsen-Scott South Pole Station, or en route to Alpha Centauri.

I feel this principle of universality is also the driving force behind wet lab biology as well.  By working in a laboratory with standardized equipment, protocols, reagents, and controls, the goal is to get reproducible results (Bruno Latour essentially agrees).  Though of course the down side is the question of whether these results have so-called external validity.  Approaching it another way, biology just may not be standardizable.

As Purnick and Weiss write in a review article about synthetic biology, “it is possible that existing engineering design principles are too simplistic to efficiently create complex biological systems and have so far limited our ability to exploit the full potential of this field.  The challenges faced in creating larger functional systems out of modules are both exciting and daunting.  The precise details of most biological environments are poorly understood.  Thus, engineering biological systems probably requires both new design principles and the simultaneous advance of scientific understanding.”

They go on to say “Standardization in other engineering disciplines allows components to be easily combined to form larger systems, but this approach relies on modularity between these components. A prevailing assumption in synthetic biology is that biological components are, or should be, modular as well. However, characterization, standardization and modularity are affected by cellular context. We cannot assume that a functional module in one cell type will work the same way in even a closely related cell type.”

These problems exist not just in biological systems, but also in human systems.  As has been noted: “But economics will forever have to contend with the biggest X factor of all: people. As Mr. Solow notes, you feed people poison, and they die. But feed them a subsidy and there is no telling what will happen. Some will use it wisely, others perversely and some a mix of both.”  Context affects people just as cellular context affects biological processes.

So with that X factor of people, is there a new hope for me going to Antarctica?  As you know, I am studying service science these days.  Moreover, some have argued that the defining feature of service  is that “Simply put, service systems have humans inside.”  Indeed, in some of the work I have been doing, I have been discovering quantifiable cultural differences between, say India and the United States.

Wouldn’t it be something if the X factor drew me into a journey of discovery that I barely survived?  Maybe I’d end up with both stories and statistics to share with you.

h1

Point Processes and Tandem Queues

November 15, 2010

Now that you have broken the pattern of alternating posts by me and you, I guess the waiting time for the next post on this blog is no longer governed by a tandem queue.  Interesting stuff, by the way.  Although analyzing the waiting times for this blog in particular is too meta, I think a general study of waiting times in social media would be very intriguing.  For example, using that to predict blog death.  Are there any good corpora of blog posts or of tweets or of some such thing?

One factor that seemingly isn’t taken into account in the most basic model of tandem queues is that with two-step interview procedures, there is attrition between the first and second steps.  That is, some people are denied or drop out before moving to the second step.  I know you aren’t a queuing theorist, but maybe you can say more about that wrinkle and how it might affect the distribution.

Speaking of point processes and being at meetings in spirit, let me also mention a presentation at the SfN annual meeting this past weekend on sampling neural spike trains.  The presentation took the ideas in this related paper and focused them specifically on the design of neuroprosthetic devices.

I know some people are waiting to hear about Antarctica, and hopefully it won’t be too much longer.  Though waiting time distributions do give positive probability to really long waits.

Incidentally, you won’t have to wait too much longer for the release of some 2010 U.S. census data, so I’m sure you’ll go crazy with that.  Somehow population data about us is more interesting than population data about birds, no?

h1

Immigrant Visa Waiting Times Addendum

November 11, 2010

Señor Kermit Romeo Erasmus, just as a quick followup to my previous post, here’s the histogram of waiting times for the CR1 visa based on the same data.  This appears very much to be the waiting time distribution for a tandem queue, with the second peak being a second interview I believe.  Clearly different types of visas have different waiting time distributions. It would be an interesting source separation problem to try to figure out all of the different immigrant visa waiting time distributions from case data that is not labeled by visa type when you have prior information that the signals are waiting times from queues.  I don’t think anyone has looked at that problem.  Even the simpler case in which the signals are single delayed gamma distributions with unknown parameters would be interesting.

Bo Zhang and I talked about possible interfaces between queueing theory and machine learning a bit last week, but unfortunately this idea didn’t come to me before our discussion.

Before signing off, let me congratulate the IBM Research team that just won the Wagner Prize at the INFORMS Annual Meeting.  (I wasn’t at the meeting in person, but was there in spirit.)