A Divide and Conquer Approach to Learning from Prior Knowledge

Abstract

A major challenge in producing large-scale simulations of the type used in ecosystem modeling is the problem of model calibration. This paper presents a method for solving a particularly dicult model calibration task that arose as part of a global climate change research project. An obvious approach to solving calibration problems is to formulate them as global optimization problems in which the goal is to nd values for the free parameters that minimize the error of the model on training data. Unfortunately, this global optimization approach becomes computationally infeasible for many real applications. This paper presents a new divideand -conquer method that analyzes the model to identify a series of smaller optimization problems whose sequential solution solves the global calibration problem. This paper argues that methods of this kind|rather than global optimization techniques|will be required in order for agents with large amounts of prior knowledge to learn ...