This is just a quick link to some photos I have taken before and during this great conference. Enjoy!
I came to San Francisco aiming to know what is going on in the field. In most part, I have tried to stretch the boundaries of what is in my comfort zone. The result has been awesome: I got aware of many interesting research work going on, of great people behind them, and also got some interesting notes for my own research agenda.
Besides that, I met in person a number of people I have been following online, like Sertalp and Pelin Çay, Marc-André Carle, and Marco Lübbecke (got all your special characters right?). Not to mention those I saw before in the 2012 meeting. We even managed to promote a sequel to the official tweetup in a local pub before the last conference reception!
Finally, participating in the chapters/fora meeting was instrumental for me and Alex Kazachkov, my colleague at CMU who lead the foundation of our student chapter. I hope we can soon put some interesting ideas we heard here in practice.
Hope to see all of you next year in Phillie!
PS: Thank you for the iPad, Gurobi!
While it’s unbelievable that the 3 rich days of the conference are coming to an end, as I look back at all the things l learnt to take back to my organization. Feels like a year long journey has just begun…
I saw the blog talking about 3 questions to be asked at the end of the conference. To build on this thought I wanted to share some specific learnings/outcomes for me from the conference
- Increase my tool box: Often most of us as analytics professionals, end up focusing so much time and energy on a specific method like in my case it was simulation, we often forget to think of using tools that might be a better fit for certain questions. The conference was a good reminer that generated ideas and thoughts around use of “unfamiliar” methods of my disctonary on the projects I am currently supporting. One specific example that comes to mind is use of system dynamics to understand the synergistic relationships/impact of PCMH model.
- New Research Questions: On the other end, the panel discussions at the annual conference were amazingly helpful. I was fortunate to participate on one organized by Health application society that discussed the challenges in healthcare and how OR can help.
- Bridge the gap between research and practice: As always presentations varied from being extremely theoretical to being extremely practice. But it was pleasing to see that gradually but surely more and more presentations/work is moving towards the center of the distribution. Researchers are thinking practical value and practitioners are recognizing the value of research driven methods. Makes me wonder the reason for this shift in the right direction – more collaborations, push from organizations like INFORMS on applied work initiatives .. whatever be the reason we are moving in the right direction and as a conference attendee, I found more value in these presentations unlike in the previous years where I would sit through the presentations and realize it was a waste of my time. Kudos to the conference committee for providing the
- My community: Best part of being at INFORMS is getting to know your community. Whether its receptions, presentations, business meetings, special invited luncheons and dinners. Even if you connect with one person at every event you can go back with wealth of know several potential collaborators and friends. At this conference I collected 20 business cards – how many friends did you make?
- Leadership: INFORMS conferences provide access to the society leadership to pitch your ideas to them. As a citizen/member of INFORMS, you have every right to share your opinions and I feel it’s my social responsibility to do so. When do you want OR/analytics to be in 2020? I was able to share my thoughts at several business meetings and in 1:1 discussions with members of roundtable, board, conference committee etc.
- Jobs/Opportunities: The career center was amazing! The Big data and analytics buzz seems to be spreading like wild fire. Although, I wasn’t actively looking for jobs, the career center gave me a taste to future career opportunities I can consider and how to navigate my way to it.
So, in summary, I am taking a lot of energy, valuable information and enthusiasm from the 3 incredible days in SFO back to Rochester, MN that will keep me busy until the next annual conference in Nov 2015. Looking forward to come back next year again for another information overload! In the meantime, it’s time to prepare for Huntington beach – abstracts for presentations/posters are being accepted NOW!
In my previous post, I covered the ways I benefit from the annual meeting. As the conference is coming to an end, I would like to wrap-up the things I would like to say before my flight to home.
Besides preparing my slides, I spent most of my time hanging around to learn more about computational stuff. I attended the COIN-OR’s developer/user lunch and heard a quite interesting discussion. It seems younger generations want to see COIN-OR’s core projects on Github, while the developers thinks that it is easy to keep the things as they are. I chit chat with Ted Ralphs and Marco Lübbecke on the issue, and Prof. Ralphs told me that he find git “too powerful” and explained briefly why such a transition will be painful.
On Monday, I joined another session (MD09) where the researchers from the SAS talked about big data optimization. It was a quite interesting session. Especially Joshua Griffin’s talk was very informative about the application and the literature. Although I am not currently working on big data, I decided to keep watch on studies on this topic.
The most hyped event of the INFORMS to me was the Computing Society Business Meeting. To our disappointment, there were no snacks this time, but the drinks and networking were great. We find an empty stroller next to the drink bar with Paul Rubin, which proves the point that people of computing society like to drink whatever their age are 🙂 Well, we expect some snacks at the next meeting! That was the main reason why human pyramid failed this time. By the way, ICS board encouraged people to record videos about their papers on INFORMS Journal on Computing. We will just have to wait and see if new OR celebrities from the ICS will emerge.
As a side note, our Lehigh University INFORMS Student Chapter got the Cum Laude award at the student award ceremony. I am proud of the chapter and our increasing involvement to the society.
At last but not least, I would like to thank you INFORMS Staff and Annual Meeting Committee for this great meeting. If it weren’t for your efforts, we couldn’t end up having this great experience. I hope to see you all next year at Philadelphia, where we will be the host. Goodbye, until next time!
What a great 3 days at the Annual Conference! I spent most of my time networking (which is THE most important thing to do at any conference), attending Decision Analysis sessions, holding the Analytics Section business meeting and participating in a panel discussion on OR and Analytics.
Some comments on each:
Decision Analysis — this is something close to my heart, as I learned decision analysis methods from a DA master at Procter & Gamble and dear friend, Joel Kahn, who recently passed away from MS. It was wonderful to see the elders of DA – Ron Howard, Jim Matheson, et al, talk about life lessons on practical applications and what they would have done differently. That alone was worth the price of admission!
Analytics Section — this year we transitioned leadership of this section from Don Kleinmuntz to myself, and along with Jim Williams (CAP Numero Uno), Polly Mitchell-Guthrie and Tarun Lal, making up the leadership of this fast-growing Section for 2015. The attendance at the business meeting was excellent — roughly 80 or so — and many engaged in several spirited conversations on CAP (which is now outpacing PMP in participation rate) and positioning of Analytics. We discussed focus areas for 2015, including increased Capabilities (CAP, Maturity Model) Visibility (communications), Recognition (awards), Conferences (Annual, Regional, International), Outreach (supporting the Analytics Ecosystem). We received great feedback from the audience on where to focus our energy in each of these areas. The good news is that based on a recent survey, 75% of the 1,065 members of the Analytics Section were willing to volunteer to help drive these efforts forward. Go Analytics!
Networking — it’s great to meet up with colleagues who you only have an opportunity to see once or twice a year, at best, as well as meeting new people. First, congratulations to the newly inducted INFORMS Fellows, and particularly to my friends Russ Labe and Radhika Kulkarni. I’ve known both for over 10 years and can attest to their passion and to their contributions to our profession. Congratulations! I also spent time meeting new connections through the “Coffee with a Member” initiative — what a great idea! Sign me up for next year!
Panel Discussion on OR and Analytics – Polly Mitchell-Guthrie did a great job recapping this in an earlier post, so I won’t repeat that here, but will reinforce that Analytics is, among many things, a way to frame the conversation with others outside of our profession. As I mentioned during the discussion, P&G “re-branded” Operations Research as Analytics over 30 years ago, for many of the same reasons why INFORMS uses the term Analytics today — to provide a “kinder, gentler” way to engage decision-makers. Thanks to Don Kleinmuntz for organizing and to Jack Levis for leading this discussion.
Along those lines, I also offer here a simple way to think about that framing. As originated by Tom Davenport and further developed by SAS and IBM and is now represented in the Analytics Maturity Model as “Descriptive, Predictive and Prescriptive”. I prefer “DAD” — Describe, Anticipate, Decide. More on that in a later post, perhaps.
Overall, a great 3 day event and I’m looking forward to the Practice Meeting in Huntington Beach. Thanks to all of the INFORMS staff for all their hard work!
I went to two panel discussions today that focused on careers in industry – the first was “What is Industry Looking for in Analytics Hires?” and focused on the demand side of this equation. This panel was formed from volunteers associated with an INFORMS committee on the teaching of business analytics, and it included Jeff Camm, The University of Cincinnati; Melissa Bowers, University of Tennessee at Knoxville; Pooja Dewan, BNSF Railway; Russ Labe, Bank of America; and Jeff Winters, UPS. This committee is trying to find out what industry wants (and has an industry subcommittee for that purpose), as well as focus on curriculum and information (essentially information-gathering and a survey of the programs), with subcommittees for each of those areas, too.
They presented great information on a survey done of CPMS, the Roundtable, and the Analytics Section on what industry wants. Then they presented on behalf of Goutam Chakraborty from Oklahoma State University, who couldn’t attend. I was sorry to miss him, because he’s a dear friend who does great work I recently wrote about in a post on 10 tips of where to find unicorns, aka data scientists. Goutam and one of his grad students used SAS Text Miner to look at job placement ads and identify the most common topics mentioned in listings. The fact that personal and communication skills came out top (65%), ranking significantly higher than the next topic, modeling and methodology (47%), generated interesting debate. Panelists from industry all agreed that the so-called “soft skills” are absolutely essential. Pooja pointed out that most PhD programs teach students strong fundamentals and require research that is done on your own, which is in contrast to the highly-collaborative nature of most industry roles, where to get anything done almost always demands working with people across other functions and divisions.
After lunch I went to the “Industry Job Panel,” which was focused on the supply side by helping students think about how to get a job in industry if that is their goal. It was facilitated by INFORMS Past President Anne Robinson, Verizon Wireless, and included Dan Fylstra, Frontline Systems; Theresa Kushner, VM Ware; Thomas Olavson, Google; and Erich Morman, a recent PhD who’d just found a new job. Some of their top advice?
- Be clear on what you want and make sure your resume reflects that focus.
- Be yourself – if you don’t fit what the company is looking for it probably isn’t a good fit for you, either.
- Do your research on the company where you are interviewing and know how they make money. Be prepared to talk about how you will add value and contribute to their ultimate goals instead of just focusing on what this role might do for you.
- In an echo of the earlier panel, be sure you listen well, show a positive attitude, and are prepared to show how you can collaborate and work on teams. Thomas said he’ll take a risk on whether you might be good or great but not on whether you have these interpersonal skills.
Finally, some students asked questions about how to answer “case” kinds of questions or illustrate practical or domain experience if you haven’t had it. The advice was to remember that these questions are rarely about whether you get the right answer but are looking to understand how you think and approach a problem. They also suggested students look for opportunities to get industry experience, whether it be internships or projects.
Which brings me to my plans for my last day of the conference. Tomorrow I’m chairing a session on “Exposing Students to Practice with a Case Competition,” which is at 8:00 a.m. and will talk about the SAS and INFORMS Analytics Section Student Analytical Scholar Competition. This provides students an opportunity to analyze a case study that is based on a real-world project and describe how they would address it in a Statement of Work. As the panelists advised, the focus of this competition is less on the single best technical solution submitted but more on the overall proposal, which includes the “math” but also the salability, understanding of business issues, etc. I will provide an overview of the competition, my colleague Jeff Day will talk about how the judges choose the winners, professor Young-Jae Jang from KAIST will talk about industry-academic collaborations and why he encouraged his student to participate, and 2013 Honorable Mention Shin-Woong Sung will talk about how this competition fit into her PhD studies and what she gained from participating. Hope to see you bright and early!
The IFORS Distinguished Lecture was given by Dr. Lyn Thomas from University of Southampton. The lecture was both insightful and interesting. Dr. Thomas first introduced the history of credit scoring, which uses operations research and statistical models to assess default risk, for consumer lending. The fact that San Francisco is the birth place of credit scoring is a happy coincidence. Among the different approaches have been used in credit scoring, logistic regression is the most common one being used now. Besides logistic regression, classification trees and ensemble models are currently in use as well. Credit scoring for subprime mortgages played an important role in the financial crisis. There were two credit models mainly used in the financial crisis: bureau scoring models and rating agencies’ models. The objective of the Fannie Mae, Freddie Mac, and rating agencies such as Standard & Poor’s was to acquire loans and securitize them. Therefore, they strongly tended to give subprime loans very high scores so that they can sell the products to investors. Dr. Thomas suggested seven lessons for O.R. in credit scoring, and I quote:
1. Ensure model objective relevant to decision maker.
- Know if static model is sufficient or if situation dynamics needs to be included.
- Making models “public” means model will be gamed.
- Verify data used in model.
- If updated data becomes available, build model to use the updated data.
- Model extrapolation can be dangerous. Credit rating agencies used corporate model for consumer CDOs.
- If model disagrees with common sense, think before using the model.
One of my favorite events of the Annual INFORMS conference is the WORMS (Women in Operations Research and the Management Sciences) luncheon at which the WORMS Award, which has been given for about a decade is announced.
This is the venue that I tell my former PhD students that they should make every effort to attend and this year the lunch sold out quickly, so the news has spread!
The great master/mistress of ceremony was Professor Susan Martonosi of Harvey Mudd College, who is serving as this year’s WORMS President. She recognized in her welcome speech that this year we have a lot to celebrate. Two former WORMS award winners, Professor Candi Yano of UC Berkeley (who is the organizer of our conference) and Dr. Radhika Kulkarni of SAS, were inducted yesterday as two of the twelve INFORMS Fellows this year and Professor Laura McLay received, this morning, the Moving Spirit Award for her amazing work with WORMS!
The lunch was delicious and especially the dessert – a chocolate bombe creation with fruit, mousse, and cake inside!
We all did the drumroll and when Professor Aleda Roth of Clemson was selected as this year’s WORMS Award winner the audience burst into applause.
The photos above were taken by Professor Min Yu of the University of Portland, so thanks! So great to see so many former students and so many friends there.
I am sure you are enjoying the conference with excellent sessions, presentations, and speakers; career fair; exhibits by vendors; social events; and networking. A big applause goes to the conference committee, but a bigger applause goes to the INFORMS staff for making it all possible and run smoothly!! Thank you!!
The plenary lecture by Morse award winner Dimitri Bertsimas was standing room only, and deservedly so. (Note to self: see if there is a polynomial time algorithm for the Geek Packing Problem.) The photo below was taken long before the arrivals stopped.
As noted by Harrison Schramm, Professor Bertsimas’s central theme was that a number of related problems in statistics (including feature selection and the development off regression models) can be viewed profitably as discrete robust optimization problems, a perspective that seems not to be in wide use in the statistics community. Professor Bertsimas also drew an important distinction, often overlooked, between computational complexity (especially P v. NP) and computational tractibility (can we solve the instances we actually have in the time and memory available to us).