The quality measures used in information retrieval are particularly difficult to optimize directly, since they depend on the model scores only through the sorted order of the documents returned for a given query. Thus, the derivatives of the cost with respect to the model parameters are either zero, or are undefined. In this paper, we propose a class of simple, flexible algorithms, called Lambd...

Standard graph theoretic algorithms are applied to chaotic dynamical systems to identify orbits that are optimal relative to a prespecified cost function. We reduce the targeting problem to the problem of finding optimal paths through a graph. Numerical experiments on one-dimensional maps suggest that periodic saddle orbits of low period are typically less expensive to target (relative to a fam...