View the complete group image at the
University of Washington program summer 2008.
|
Andrew Becker
Featured Scientist Interview
|
|
In a recent analysis of
Essential Science IndicatorsSM data from
Thomson
Reuters, the work of Dr. Andrew Becker entered the
top 1% in the field of
Space Science with the highest citation count. His
record in this field includes 63 papers cited a total of
2,428 times between January 1, 1999 and December 31,
2009.
Dr. Becker is a Research Assistant Professor in the
Department of Astronomy at the University of Washington.
Below, he talks with ScienceWatch.com about his
highly cited work.
|
|
Would you tell us a bit about your
educational background and research experiences?
My interests in Physics and Astronomy were piqued somewhat randomly around
7th-8th grade when I picked up a used book by someone I had occasionally
seen on TV named Carl Sagan. The book was called Contact and dealt
in a very fascinating way with the nature and pursuit of knowledge, framed
by this awesome circumstance of being contacted by extraterrestrial
intelligence.
The characters really resonated with me, so I naturally became curious
about the subjects they pursued in the book, namely mathematics but also
the interplay between science and religion on an intellectual level. I was
recently honored to meet Jill Tartar, on whom the protagonist Ellie Arroway
was based, this past year in Washington, DC.
Amusingly, at that time I was also interested in books by another person I
recognized from TV, L. Ron Hubbard, although I did not venture far from his
popular science fiction books. My life certainly might have turned out
differently had I ventured into Scientology instead of the physical
sciences!
I decided to major in physics in college after experiencing a very
invigorating AP Physics course in high school. As an undergraduate, I
followed the well-trodden path of searching for basic jobs to sustain
myself, eventually starting work in the undergraduate labs resetting the
frictionless track experiments after hours.
After demonstrating a general sense of responsibility, and having done
relatively well on course work, I started working on a cosmic ray
experiment with Dr. Jim Gaidos, learning about photomultiplier tubes and
turning flat glass into mirrors in the basement tunnels of Purdue
University, right next to their tandem accelerator. We also worked on
repairing the local Cumberland Observatory, including basic upkeep such as
removing families of dead spiders, metalworking on the observatory dome,
the purchase of eyepieces, and eventually undertaking successful
observations of the Shoemaker-Levy 9 impact with Jupiter.
Here, I learned the basics of measuring the brightness and positions of
stars in astronomical images. I found I had an aptitude for the necessary
computer programming, originating from my days programming our TRS-80
computer, when I saved my work on cassette tape and worked on modifying the
game algorithms to allow me to win. This programming background, along with
the typing class my Mom made me take in high school, are two of the core
foundations that have since allowed me to operate effectively in my field.
|
"Because of advances in hardware, software, and
information technology the next decade of astronomy
promises a deluge of information."
|
I moved to Seattle for graduate school, impressed by the natural setting of
the city and the overall kindness of the students and department. The day I
walked in the door I was offered a summer job working on one of the first
large-scale time domain astronomical surveys, the MACHO microlensing
project, coordinating global follow-up of events we had found.
It was here that I learned about writing general solutions to problems;
your algorithms need to be robust if they are being applied to multiple
data sets, each with its different kinks and wrinkles. I learned many of
the benefits and complications of working within a large collaboration.
Some of the team meetings were intense, even tearful, but learning the
dynamics of discourse in large scientific groups has helped me to feel
comfortable in the many collaborations I have joined since.
The needle-in-a-haystack nature of the project also required that I start
to come to grips with some of the difficulties faced by astronomers today:
anomaly detection in huge datasets, classification of events, coordinating
global resources to follow-up events, and uses of advanced fitting
techniques to compare models to data.
My first post-doctoral position was at Bell Laboratories, the legendary
research institute where the transistor was developed, the C programming
language and UNIX were written, and where the cosmic microwave background
was accidentally discovered by laboratory scientists, ushering in a new age
of experimental cosmology. While here I came to appreciate the benefits of
having the intellectual freedom to pursue what interests you, and to be
grateful for having the opportunity to do so myself.
The folks at our lunch table seemed to compete with each other to come up
with the most extraordinary ideas for patents, such as using voltage spikes
on undersea cables as early warning signs of tsunamis. I was somewhat
embarrassed by the narrowness of my knowledge in only one area of the
physical sciences; it was extraordinarily humbling.
It was here that I started to develop generalized variability detection
code that used the simple premise that anything left in the
difference between two images represents something that has varied
in brightness or position. This software has since been made exceedingly
robust by being applied to various time domain searches for distant
supernovae, dark matter microlensing, Solar System asteroids, and to other
generalized surveys measuring the distribution of the matter in the
Universe.
This software became the catalyst for my involvement in the many
astronomical collaborations that have led to my apparently prodigious
citation record: the MACHO project, Microlensing Planet Search, Deep Lens
Survey, ESSENCE survey, SuperMACHO survey, SDSS-II Supernova Survey, and
Large Synoptic Survey Telescope (LSST) Collaboration.
I have since returned to and am now research faculty at the University of
Washington in Seattle, where we are busy designing the next generation of
optical astronomical time-domain surveys, the Large Synoptic Survey
Telescope (LSST).
It's clear from reading the above I have been given many opportunities to
succeed. I try and afford these opportunities to other young students
through my involvement with various outreach programs, including the
University of Washington's Pre-Major in Astronomy Program (Pre-MAP) and
through the National Science Foundation's Faculty and Student Teams (FaST)
program.
What would you say is the main focus of your research?
The vast majority of my research involves searching for time-variable
phenomena in large astronomical data streams. This variability may
originate from objects relatively close to home, such as moving asteroids
orbiting our own Sun, or may be pulses of light that come from exploding
stars in the distant Universe. The intent is to build a common platform
that can measure and distinguish between these different classes of
phenomena.
This involves addressing many difficult problems along the entire path from
telescope to knowledge inference: data acquisition and storage, data
reduction and mining techniques, designing statistical models of the data,
and drawing conclusions from comparison of these models to the data. It
also requires large collaborations of smart and competent scientists, all
of whom have contributed greatly to my publication record, and whom I must
thank for being recognized here.
Your most-cited paper, and several other of your highly cited
papers, deals with the MACHO project. Would you talk a little about this
aspect of your work?
The MACHO project was one of the first large-scale, real-time, optical
astronomical time-domain surveys. We were searching for (literally)
one-in-a-million stellar brightenings caused by gravitational lenses along
our line of sight to background stars.
We had to study dense stellar fields, such as towards the Galactic center,
to have any chance of seeing even one of these events. We also had to
marshal resources to study these events once they were found.
In many ways it set the stage for my work today, which is the application
of these principles on a much larger scale. I also met many of my current
collaborators from working on MACHO. It was a very successful project in a
variety of ways, and was a great opportunity for a young graduate student.
The ESSENCE Supernova Survey also turns up quite a bit in your
highly cited papers. Would you talk about this project and your involvement
in it?
At the time of ESSENCE's inception, the High-Z and Supernova Cosmology
Project collaborations had recently announced the first detection of "dark
energy" (Riess et al. 1998; Perlmutter et al. 1999), a
mysterious component of the Universe that is causing its expansion to
accelerate.
In these surveys, time-variable supernovae were used as bright beacons to
probe the expansion history of the distant Universe, a groundbreaking task
that will almost certainly lead to the Nobel Prize for members of those
teams.
We had recently started a successor to MACHO called (naturally) SuperMACHO,
and I was working in Chile building an automated image subtraction pipeline
with my invaluable collaborator Armin Rest.
While we were writing the code, the project's principal investigators were
working on a parallel idea: we would use the same telescope and same
software in a targeted search for supernova events instead of microlensing
events! It was a challenge to make our software robust enough to operate in
both regimes, but inevitably made the pipeline that much more valuable.
So the ESSENCE project is a nifty example of the notion that all problems
are interesting and many of their solutions are interconnected. If you can
make that connection, intellectually and programmatically, it can lead to
new collaborative efforts.
Of all the projects on which you have worked, do you have a
particular favorite?
My favorite era was from 2000 to 2003, when I was spending many months down
in Chile working on the Deep Lens Survey, SuperMACHO Microlensing survey,
and ESSENCE Supernova survey. This was the time of my post-doc at Bell
Labs, and while I'm not sure they were appreciative of all the time I spent
in South America, it did give me the intellectual freedom to get something
very important done—to start facing some of the challenges that will
come to define the upcoming decade in time-domain astronomy.
In some sense this is exactly what is supposed to happen during your
post-doctoral tenure, before you become overwhelmed with the academic and
teaching responsibilities of being a professor. Much of the software
written during this era in still in use today, some of it even serving as a
reference for the next generation of astronomical applications.
Are there any projects you have forthcoming that you are free to
discuss?
The most visible is the Large Synoptic Survey Telescope. This survey
promises to acquire, archive, and serve to the general public more data
than has been taken in the prior cumulative history of Astronomy! It
will survey the entire visible sky every three nights, continuously for
10 years. Most importantly, we will be releasing these data to the
general public (amateurs as well as professional astronomers) in truly
real-time.
|
"The vast majority of my research involves searching
for time-variable phenomena in large astronomical data
streams."
|
This is the democratization of science, simultaneously enabling research by
professors at universities with limited resources, as well as by faculty at
universities with access to the largest telescopes on the planet. The
results of this "cosmic cinematography" will also serve as a valuable
teaching resource for education and public outreach.
From a technical point of view, the data challenges here are enormous. We
need to mine these data in real-time for outstanding events that deserve
additional and immediate study. To do this requires we recognize and
classify the more prosaic types of variability that may masquerade as these
rare phenomena.
To maintain the community's trust we need to do this with an appropriately
low false-alarm rate. All the aspects of this problem require that I
broaden my own knowledge and understanding of statistical modeling,
classification, and data-mining techniques.
In what directions do you see the field of astronomical surveys
going in the next decade?
Because of advances in hardware, software, and information technology the
next decade of astronomy promises a deluge of information. Much of the
science will not get done at the telescope, but instead by an astronomer
sitting in front of a computer and executing queries on a massive database.
This requires an entirely different skill set than has been needed to excel
in astronomy in the past, and thus it's imperative that we start to teach
these skills to graduate students as soon as possible.
However, optical astronomical surveys such as LSST are only one facet of
the next decade of astronomy. Due to new technologies we are also probing
the Universe in fundamentally new regimes, from gravity waves to neutrinos.
We are also slowly but surely developing the technologies needed to search
for the signatures of planets around other stars, and for ascertaining
whether or not they may harbor life. This includes direct imaging
techniques, methods to infer the atmospheric properties of the planet, and
even searches such as SETI looking for signatures of technology coming from
these distant planets.
The challenge in all these cases is to find the needle in the haystack,
that rare event that changes our understanding of the field, but whose
discovery is enabled by the accumulation of massive amounts of data.
Acquiring and understanding these data pose the largest challenges for our
field in the next decade.
Andrew C. Becker, Ph.D.
Department of Astronomy
University of Washington
Seattle, WA, USA
Andrew Becker's current most-cited paper in Essential Science
Indicators, with 322 cites:
Alcock C, et al., "The MACHO project: Microlensing results from
5.7 years of Large Magellanic Cloud observations," Astrophys. J.
542(1): 281-307, Part 1, 10 October 2000. Source:
Essential Science Indicators from
Thomson Reuters.
Additional information:
KEYWORDS: ASTRONOMICAL SURVEYS, MACHO PROJECT, MICROLENSING, VARIABILITY
DETECTION CODE, MICROLENSING PLANET SEARCH, DEEP LENS SURVEY, ESSENCE
SURVEY, SUPERMACHO SURVEY, SDSS-II SUPERNOVA SURVEY, LARGE SYNOPTIC SURVEY
TELESCOPE (LSST), TIME-VARIABLE PHENOMENA, LARGE ASTRONOMICAL DATA STREAMS,
COLLABORATIONS, DATA MINING.
