Institute for Mathematical Sciences                                        Programs & Activities



Online registration form



      Scientific aspects


Workshop on New Directions in Stein's method
(18 - 29 May 2015)


Organizing Committee · Visitors and Participants · Overview · Activities · Venue


 Organizing Committee



 Visitors and Participants





One of the greatest accomplishments of probability theory is its success in expressing specific aspects of complicated random phenomena by means of relatively simple limiting distributions. These limits often exhibit a certain “universality” in that they depend only on the most fundamental properties of the models of interest. The classical central limit theorem is the prototype of such results, stating that, under some rather weak moment conditions, a sum of independent random variables, normalised appropriately, will converge to a Gaussian distribution. With approximations of finite systems by their limits, both in theory and applications, comes the need to bound the approximation error, and it is here where Stein's method, introduced in 1970 by Charles Stein, has played an important role over the last few decades.


Perhaps the main advantage of the method is its robustness with respect to dependence in probabilistic models. The method is by no means confined to sums, or to independent random variables, or to approximating distributions such as the normal, which has a nice and explicit probability density function. Indeed, Stein’s method for both normal and non-normal approximations has found applications in a large variety of areas, including statistics, point process theory, combinatorics, number theory, random graph theory, random matrix theory, and statistical physics, to name but a few.


The past two programmes in 2003 and 2008, and a workshop in 2009, all held at IMS, illustrate both the variety and the depth of the field, which has continued to grow since then. The combination of Malliavin calculus and Stein’s method has proved to be very fruitful, having been applied not only to functionals of Gaussian fields, but also to functionals of Rademacher sequences and Poisson point processes. The use of Stein couplings to prove concentration of measure inequalities has opened a new avenue for handling complicated systems exhibiting genuine dependence. Stein’s method has also been extended to obtain very refined moderate deviation results. Multivariate normal approximation, technically notoriously difficult, has seen important developments, for example pushing the dependence on the dimensionality towards the optimal range. Various new distributional transforms have been discovered and successfully combined with Stein’s method.


Considering the diversity and the new exciting directions of recent advances, we believe now is the right time to organise a comprehensive workshop on Stein's method and its applications. We plan to bring together not only active researchers directly working in the area, but also those who apply Stein's method in their work in order to stimulate, strengthen and develop existing interactions between theory and practice. The National University of Singapore has long been a centre of research in Stein's method, having been the base of major programmes and the home to a number of researchers who have been influential in its development. The combination of these many factors makes Singapore a natural place in which to conduct the workshop, with the resulting benefit of fostering collaborations, both locally and internationally.


  • Workshop, 18 - 29 May 2015


Students and researchers who are interested in attending these activities are requested to complete the online registration form.

The following do not need to register:

  • Those invited to participate.



Organizing Committee · Visitors and Participants · Overview · Activities · Venue

Best viewed with IE 7 and above