Nancy L. Geller

Nancy Geller with plenary speaker C. R. Rao (left)

As 2011 ASA president, I was invited to be a plenary speaker at the International Statistics Conference last December in Columbo, Sri Lanka. The other plenary speaker was C.R. Rao, and I was quite honored to be in his presence.

The conference, themed, “Statistical Concepts and Methods for the Modern World,” took place December 28–30 at a state-of-the-art conference center on the outskirts of Columbo. The ASA was a silver sponsor for the conference, which is why I was invited to present a plenary talk.

The conference began with a traditional opening ceremony. The vice chancellor of the University of Columbo and the Sri Lankan Senior Minister of Science and Technology were present. One by one, all special guests (including me) lit the opening torch. Then came the melodious national anthem and the presentation of an honorary degree—his 34th—to C. R. Rao. Rao’s degrees are from 18 countries and include the Guy Medal (Britain), membership in the U.S. National Academy of Sciences, and the U.S. National Medal of Science. Rao also was awarded the first fellowship of the newly created Institute of Applied Statistics of Sri Lanka. Making the Sri Lankan Applied Statistical Association into an institute took an act of parliament, a recent achievement of its president, Palitha Sarrukali, who was one of my hosts.

Chief guest and Sri Lankan Senior Minister of Science and Technology Tissa Vitarana spoke about global warming through human intervention and its potentially disastrous effect on island nations. He urged making policymakers and the general population more aware of the magnitude of the problem and urged statisticians to contribute to this effort. He spoke about statisticians’ contributions to modeling after the 2004 tsunami and expressed optimism that this meeting would strengthen applied statistics and its contributions to scientific research.

The first plenary talk by Rao was titled “Importance Sampling and Cross-Entropy with Applications to Problems of Large Deviations and Optimization Problems.” Cross-entropy is a measure of loss of information when one uses a probability density g to make inference on a probability density f. One chooses g by minimizing cross-entropy and importance sampling, and then one can use Monte Carlo techniques to find probability of large deviations or the maximum of a function over a given subset.

I gave the second plenary talk on the morning of the second day. The talk, on prognostic and predictive biomarkers, was well received with questions and a number of inquiries from students who want to study or do post-docs in the United States. Most of the talks were mathematical statistics, and mine was more applied.

Overall, the conference was interesting with a wide range of topics, such as robust statistics, Bayesian methods, cancer research, and sequential analysis.

I also gave a seminar at the department of statistics at the University of Columbo. Attendance was better than at most talks I’ve given in the United States, as many former students came ‘home’ for the end of the year break and to attend the conference the following week.

Sri Lanka is just north of the equator, and although December is winter, it was hotter during the day than in Miami in August! The Sri Lankan statisticians were warm and welcoming. The ASA has made many new friends.