Clearing the Jungle of Stochastic Optimization
Abstract
Stochastic optimization for sequential decision problems under uncertainty arises in many settings, and as a result as evolved under several canonical frameworks with names such as dynamic programming, stochastic programming, optimal control, robust optimization, and simulation optimization (to name a few). This is in sharp contrast with the universally accepted canonical frameworks for deterministic math programming (or deterministic optimal control). We have found that these competing frameworks are actually hiding different classes of policies to solve a single problem which encompasses all of these fields. In this talk, I provide a canonical framework which, while familiar to some, is not universally used, but should be. The framework involves solving an objective function which requires searching over a class of policies, a step that can seem like mathematical hand waving. We then identify four fundamental classes of policies, called policy function approximations (PFAs), cost function approximations (CFAs), policies based on value function approximations (VFAs), and lookahead policies (which themselves come in different flavors). With the exception of CFAs, these policies have been widely studied under names that make it seem as if they are fundamentally different approaches (policy search, approximate dynamic programming or reinforcement learning, model predictive control, stochastic programming and robust optimization). We use a simple energy storage problem to demonstrate that minor changes in the nature of the data can produce problems where each of the four classes might work best, or a hybrid. This exercise supports our claim that any formulation of a sequential decision problem should start with a recognition that we need to search over a space of policies.
17:30
Social Capital and Microfinance
Abstract
Some density results in number theory
Abstract
I will describe joint work with Manjul Bhargava (Princeton) and Tom Fisher (Cambridge) in which we determine the probability that random equation from certain families has a solution either locally (over the reals or the p-adics), everywhere locally, or globally. Three kinds of equation will be considered: quadratics in any number of variables, ternary cubics and hyperelliptic quartics.
Mathematical modelling of epithelial dynamics: from cells to tissues
Cardiac Physiology, Theory and Simulation in the Clinic
Abstract
Computational models of the heart have been primarily developed to simulate, analyse and understand experimental measurements. Increasingly biophysical models are being used to understand cardiac disease and pathologies in patients. This shift from laboratory to clinical contexts requires the development of new modelling frameworks to simulate pathological states that invalidate assumptions in existing modelling frameworks, work flows to integrate multiple data sets to constrain model parameters and an understanding of the clinical questions that models can answer. We report on the development and application of biophysical modelling frameworks representing the cardiac electrical and mechanical systems, which are currently being customised for modelling cardiac pathologies.
Theory of evolutionary couplings and application to the prediction of protein 3D structure and fitness
Abstract
Genomic sequences contain rich evolutionary information about functional constraints on macromolecules such as proteins. This information can be efficiently mined to detect evolutionary couplings between residues in proteins and address the long-standing challenge to compute protein three-dimensional structures from amino acid sequences. Substantial progress on this problem has become possible because of the explosive growth in available sequences and the application of global statistical methods. In addition to three-dimensional structure, the improved analysis of covariation helps identify functional residues involved in ligand binding, protein-complex formation and conformational changes. We expect computation of covariation patterns to complement experimental structural biology in elucidating the full spectrum of protein structures, their functional interactions and evolutionary dynamics. Use the http://evfold.org server to compute EVcouplings and to predict 3D structure for large sequence families. References: http://bit.ly/tob48p - Protein 3D Structure from high-throughput sequencing; http://bit.ly/1DSqANO - 3D structure of transmembrane proteins from evolutionary constraints; http://bit.ly/1zyYpE7 - Sequence co-evolution gives 3D contacts and structures of protein complexes.