Takeaways from the Bayesian Nonparametrics Workshop

The first entry is from Chetkar Jha, PhD Student at Missouri University.

Group shot at SAMSI's Bayesian Nonparametrics Workshop

Attendees at the SAMSI Bayesian Nonparametrics workshop.

A couple of weeks back, I attended a workshop on “Bayesian Nonparametrics” organized at Statistical and Applied Mathematical Sciences Institute (SAMSI) .

It was a 4-day-long workshop on Nonparametric Bayesian. The goal of the workshop was to brainstorm on some of the pressing problems related to Nonparametric Bayesian and discuss possible solutions as a group. Let me describe the format of the workshop to give you some flavor. Each day was divided in two halves: morning session and afternoon session. In the morning session there were presentations on Nonparametric Bayesian and that would lead to brainstorming sessions on related problems in the afternoon session. Since, we’re working in smaller groups that gave us a lot of latitude to discuss the topics closely and ask a lot of questions and clarifications. I, for one, really enjoyed talks and discussions on convergence/contraction, variational inference, MCMC methods and scalable models. Being a graduate student, there was a lot of new content for me and it was harder to assimilate but the workshop gave me exposure to lot of new content and some topical problems.

Person talking at the Bayesian Nonparametric workshop

Interesting lectures were presented.

The workshop was attended by some of the leading researchers in the field. It was sort of a ‘fanboy’ moment for me, as I was only aware of their names and their work. The workshop provided a perfect opportunity to meet ’real’ people behind the names. Also, I loved the energy and the passion that the group shared for Non Parametric Bayesian that was really motivating and hopefully, some of it did get rubbed on me.

Also, I would like to take this opportunity to thank the organizers and people at SAMSI, who did a wonderful job in organizing the entire event. Hopefully, we can have more such workshops in the future.

The second entry is from Dootika Vats, PhD Student in the School of Statistics at the University of Minnesota

My build up to the 4th of July weekend turned out to be a rather educational experience. I was fortunate enough to attend SAMSI’s workshop on “Bayesian Nonparametrics: Synergies between Statistics, Probability and Mathematics” from June 29th to July 2nd. This was my first visit to SAMSI and to the Research Triangle area. The first thing that stands out about the area is how green it is! Calming stretches of green fields and trees, make for an ideal research environment.

Driveway with grass and trees.

Driveway to SAMSI’s building in RTP.

The 4-day workshop followed the 10th Conference on Bayesian Nonparametrics held in Raleigh from June 22-26. Many participants of the workshop had attended both events, which made the workshop a great platform to discuss key points and ideas that came out of the conference.

The workshop was attended by professors, postdocs and graduate students from all over the world. We were a small group of people that came with varied research focuses to contribute to/learn about Bayesian nonparametrics. The days were packed into discussion style seminars in the morning, followed by a delicious lunch spread, and breakout groups in the afternoon. Each day had a somewhat broad, yet specific focus of interests like multi-resolution methods, high dimensional analysis, scalability and optimization, and theoretical developments.

Food at the SAMSI workshop

The food at the workshop was splendid!

The breakout groups really made this workshop different from other conferences and programs I had attended before. Each group was led by an expert in the field, and the audience could choose any group that appealed to them. Most groups ended up with 5-10 people at most. This made for an extremely educational experience for a graduate student such as myself. We got an insight into how experts in the field approach a problem and attempt to come up with plausible solution paths. Just observing these world-class researchers openly think about a problem and having the opportunity to ask trivial questions was worth the trip!

Apart from reading an introductory paper, I was not very familiar with Bayesian nonparametrics. My research is on Markov chain Monte Carlo(MCMC) algorithms so, of course, there were times when I did not quite understand the questions put forth in discussions or the even the problem at hand. However, since there were so many young researches, post-docs and new faculty, it made it easier to ask “stupid” questions. The workshop also held a poster session for young researchers to talk about their own research. I was able to present my work on MCMC output analysis and discuss ideas and improvements over delicious food and drinks.

SAMSI Bayesian Nonparametric poster session

People talking at the poster session.

Overall, I think SAMSI put forth a wonderfully organized workshop. I came back with a better understanding of Bayesian nonparametrics and with feedback and ideas for my own research. The logistics of the workshop were also well managed with frequent communications from the staff about the schedules. And, of course, the almost endless supply of coffee was deeply appreciated! I will definitely keep a lookout for more SAMSI events and encourage other graduate students to apply for such workshops and conferences.

Postdoc Profile – Christopher Strickland

Christopher Strickland on a hill with the ocean in the background

Christopher Strickland hiking in New Zealand.

SAMSI postdoctoral fellow, Christopher Strickland was born in Houston, Texas and lived briefly there and in Dallas before he could really remember either place. He grew up in Oxford, Mississippi. His grandfather was Chair of Modern Languages at the University of Mississippi and helped to establish a study abroad program, and his grandmother was originally from France, so many summers his father and grandfather traveled to France. When Christopher went to ” Ole Miss,” in the honors college, he minored in physics before switching degrees and getting a double degree in Mathematics and French.

At first he was following a more pure math route. He went to the University of Florida in Gainesville for his Master’s degree and was studying logic, but after about a year and a half, he realized this was not the area he preferred. After changing his focus to dynamical systems and defending his Master’s thesis, he stayed in Gainesville for a year as he tried to figure out what to do next and taught mathematics at Santa Fe Community College. He knew he would prefer to get into an area that involved applied math instead of pure math. He became interested in mathematical ecology and had heard that Colorado State University had a great program in ecology and the natural sciences, so he applied there to get his mathematics Ph.D.

Christopher considers Patrick Shipman, who was a new faculty member at Colorado State at the time, and Gerhard Dangelmayr, who is the Chair of the department, to be his mentors. They were also his co-advisors. Christopher and Patrick started collaborating on projects right away.

“I was headed toward dynamical systems which is really related to mathematical ecology, so I worked with Patrick and Gerhard for the next six years,” Christopher said, “I still collaborate with both of them, and we are currently applying for a research grant to work with the U.S. Fish and Wildlife Commission.” Christopher has also collaborated with Patrick Shipman and Snehal Shetye at Colorado State on a project modeling the mechanical properties of spinal cords.

Nate Burch told me about SAMSI originally,” said Christopher. “Nate and I were colleagues at Colorado State.” So, when the Ecology program was announced, he applied and was accepted.

Christopher Strickland standing on a rocky edge

Christopher Strickland hiking in Australia.

While he’s been at SAMSI, Christopher has worked on getting various parts of his dissertation re-written into smaller parts so that he can publish each part in various journals. He has three of the four published. The manuscript of the fourth one is completed, and has been submitted as of June 2015.

Christopher has been participating in two working groups this year: The Tipping Point group and the Physical Ecology group. The Physical Ecology group led by Laura Miller, has been particularly interesting for him. “We recently had this really great workshop at SAMSI, which was for the people participating in the working group. We invited Nadia Kristensen from the University of Queensland who brought in all this great data from parasitoid wasp release and spread. That’s been really nice because I mostly do modeling of dynamic systems and the model that she had with this data could be something I could help her improve,” he commented.

“We are also working on a review paper, which is something the working group conceived of sometime around December. The entire working group and even some other people, including some ecologists and my advisor from Colorado State, Patrick, is working on this review,” Christopher said. He believes the review will be completed by the end of this summer.

Much of Christopher’s research focuses on networks, specifically looking at spread and control of contagions on the network. One example would be to look at container shipping networks or airline networks. He is working on a grant that is looking at white nose bat syndrome that involves a network of caves. While bats could spread the disease themselves from cave to cave, there is also the concern that hikers or cavers could get the fungus on their boots and spread the disease when they hike in a different cave. By figuring out how these networks work, it may help ecologists figure out where the disease might spread next, or help them to get a disease under control.

Christopher Strickland makes a kick

Christopher practicing Cuong Nhu.

When Christopher is not at work, he is either playing a game of soccer (he used to be on a math league!) or he is practicing the art of Cuong Nhu, (meaning hard/soft in Vietnamese) a type of martial arts that was brought to the United States in Gainesville, Florida. Christopher is on target to get his black belt, probably in about a year. “A lot of scholarly people actually do this type of martial arts. It has been a good way to network,” quipped Christopher. He also spends time with his girlfriend, Anne Ho, who is a theoretical mathematician. They like to travel a lot, many times to national parks or overseas.

In the fall, Christopher will be teaching at the University of North Carolina at Chapel Hill while he completes his second year as a postdoctoral fellow for SAMSI.

Why you should attend the SAMSI Forensics 2015-2016 opening workshop

The following was written by Dr. Clifford Spiegelman, Distinguished Professor of Statistics at Texas A&M and one of the program leaders for the 2015-2016 SAMSI Program on Statistics and Applied Mathematics of Forensic Science.

Cliff Spiegelman

Dr. Clifford Spiegelman

Imagine having a nightmare where nearly all evidence presented in courts was seriously misrepresented. No, not a nightmare about someone accused of being a witch, but a more current trial. Say the defendant is accused of rape or murder and all the scientific evidence presented was seriously misrepresented and biased toward the prosecution. It would not be a pleasant dream, but it is today’s reality, and that is worse than a nightmare as it is real. Within the last months the FBI has admitted to over representing the importance of hair matches for decades. Prior to that in 2007 CBLA or comparative bullet lead analysis was another procedure used for decades where the FBI admitted to overstating the importance of a match.

Forensic science is inherently a field that uses data (patterns, pictures, etc.) to link suspects to crimes. Unfortunately, the use of formal statistical methods or even statistical or mathematical thinking is uncommon.

That is where you can help. There is a dearth of persons, as in way to few mathematical scientists, that are aware of the issues.

What are the issues?

Well one can read the summary of the 2009 NRC report “Strengthening Forensic Science in the United States: A Path Forward” to get a good overall view. Here are some of my recent consults: A defendant was charged with indecent contact with a minor. The minor had chlamydia but the defendant did not and was not treated for chlamydia. What is the probability? In another case a convict has been in jail for 40 years largely based upon hair and fiber evidence. The hair evidence was inconclusive. That is the crime lab hair examiner testified that there were both similarities and dissimilarities between the pubic hairs found at the scene and on the defendant. Subsequently some inconclusive results (not the case in question as the evidence has gone missing) have been investigated using DNA. What are the odds that an inconclusive microscopic hair analysis has a DNA analysis that excludes the defendant? It is more than ½.

The opening workshop will look at various forms of traditional pattern evidence. These include fingerprints, firearm/toolmarks, shoeprints etc.. Help become part of the birth of taking forensic science from oxymoron state to a real science.

The opening workshop program can be found here. Read more about the overall program here, and if you want to learn more about forensics before the opening workshop, consider attending a special tutorial a few days before the big event begins.

Please join us. You can make a difference to the legal system and make our country a more just place.

Impressions from the Undergraduate Workshop on Data-Driven Decisions in Healthcare

big group of students outside SAMSI

February 2013 Undergraduate Workshop participants.

SAMSI recently held the Undergraduate Workshop on Data-Driven Decisions in Healthcare for about 30 students. Visiting professors, postdoctoral fellows and graduate fellows who are participating in this SAMSI program led the sessions providing cutting-edge research into the lectures. Students had a chance to work with data from the SEElab at Technion in Israel, got an overview of personalized medicine and a tutorial in R and a demonstration of the ARENA software.  Here are a few of the students’ impressions from the workshop.

Eric Laber instructing students

Eric Laber, NCSU, giving lecture at the workshop.

Eric Kernfeld, Tufts University Class of 2014, Applied Mathematics

“I had a great time at the workshop on Data Driven Decisions in Health Care this past weekend. It was a nice opportunity to meet statisticians, something I don’t get the chance to do back at Tufts. I also met a lot of undergraduates majoring in statistics and mathematics. The food was good, the staff were welcoming, the accommodations were convenient, and the talks were well-pitched. I recommend SAMSI workshops to anyone who’s interested in the topics, especially to people considering graduate education down the road.”

Danielle Llanos, Georgetown University

“I thought the SAMSI workshop was wonderful. It was a great opportunity to learn from talented individuals, and a chance to expand my network. The lecture topics were incredibly interesting and were very relevant to my career goals. Probably the best part of the workshop was the graduate student panel. The ability to ask those burning questions and learn from the experiences of others was great. I would recommend any SAMSI workshop to students looking to learn more about opportunities in the sciences, and expanding their educational experiences.”

three students at table

Students networking at lunch.

Brittany Boribong, sophomore, biomathematics major at University of Scranton

“As a student with no background in statistics and programming, I found the workshop a bit overwhelming but no less interesting. Coming into this with no experience just allowed me to take that much more out of the workshop.  I was able to explore new fields of math that I never considered before and learn about topics that I had no idea even existed. As a Biomathematics major, I found the topic of using data to derive decisions in healthcare intriguing since it is an application of my major that I was not aware of. Another wonderful aspect of the workshop was the chance to speak to people in different fields. During lunch, I had the opportunity to speak to a post-doc fellow and during dinner, I spoke to one of the professors that gave a lecture earlier in the day; these opportunities don’t come along every day. It was enjoyable hearing their stories and being able to have a casual conversation with them. The panel made up of current graduate students and post-docs was also helpful in that they were able to share their experiences about graduate school and offer along any advice. I found it particularly helpful since one of the speakers was currently in a biomathematics program and I was able to ask questions I had about my major.

However, the best part of the workshop, in my opinion, was being to meet other students. Coming from a university with a smaller math department, I really enjoyed meeting students from around the country with interests similar to my own. It was great being able to make connections with students in different fields and from universities from all over. Overall, I had a wonderful time meeting new people and exploring different fields of mathematics during the workshop and found this to be a great experience.”

The Undergraduate Workshop Focusing on SAMSI Computational Methodology for Massive Datasets

This blog entry was written by James Anderson, undergraduate student double majoring in statistics-mathematics and economics from the University of Connecticut.

The undergraduate workshop attendees

Attendees and some presenters from the SAMSI undergraduate workshop held October 26-27, 2012.

This undergraduate workshop was notably different from my previous experience, though in no way inferior.  In fact, I would argue the content of this workshop was better for my current position. Massive datasets are surprisingly common and the topics covered included astronomy, high dimension regression, climate change, and image rescaling. In these contexts, we mainly discussed how to manage large datasets without crashing an individual computer.

The other aspect of the workshop, which I really enjoyed, was discussion panels. The students got a chance to talk to people working in academia and industry, as well as graduate students and postdocs. The professionals talked about their respective occupations and how they got to where they are, which was very interesting. On the other hand, the younger group talked about their transitions out of their respective undergraduate programs. This was particularly useful as I will be going through this phase over the next few months. One thing I was once more impressed with was SAMSI’s concern for the attendees. The presenters were happy to go into great detail about their presentations and field any general discipline related questions they could with interested attendees (the presentations had to be kept pretty short). This really impressed me; it didn’t matter if it was in the context of a presentation or not, the mentality seemed to be that the workshop was happening all the time. There was a great opportunity during panels or breaks to ask questions and get information that was quite personalized and would have been hard to find in another way. The workshop gave me a lot of information and resources that will be valuable going forward.

Apply Now for the 2-Day Undergraduate Workshop at SAMSI October 26-27

group of undergraduate students from 2011

Last year’s undergraduate workshop group.

SAMSI is accepting applications for the two-day undergraduate workshop that will focus on Statistical and Computational Methodology for Massive Datasets. The workshop will be held October 26-27 at SAMSI in Research Triangle Park, NC. The program begins at 9:30am on Friday, October 26 and ends at noon on Saturday, October 27.

Applications received by Friday, September 28 will receive full consideration. SAMSI will reimburse appropriate travel expenses, including food and lodging. Participants are urged to arrive on Thursday evening.

The Statistical and Computational Methodology for Massive Datasets program focuses on fundamental methodological questions of statistics, mathematics and computer science posed by massive datasets, with applications to astronomy, high energy physics, and the environment. Serious challenges posed by massive datasets have to do with “scalability” and “data streaming.”

Computational Advertising Summer Program

The SAMSI computational advertising summer program is heating up! Last week, the group met at the Radisson RTP and had a great series of lectures from both professors at various universities from around the country to some of the cutting edge research happening at companies like LinkedIn, Yahoo!, Facebook, AT&T and Maxpoint Interactive.

Christian Posse from LinkedIn before speaking at the workshop

Christian Posse from LinkedIn talked to the people attending the computational advertising summer program at SAMSI.

Some of the people who spoke last week have shared their lectures and are available to view on the SAMSI website.

Zainab Jamal from HP presenting her poster

Zainab Jamal from Hewlett-Packard sharing her poster at the Computational Advertising poster session.

Several of the participants of this program will be sharing their experiences with you in the next few days. Stay tuned for more on this subject!