A modification of the standard Simulated Annealing (SA) algorithm is presented for finding the global minimum of a continuous multidimensional, multimodal function. We report resu...
A decision process in which rewards depend on history rather than merely on the current state is called a decision process with non-Markovian rewards (NMRDP). In decisiontheoretic...
In many application domains there is a large amount of unlabeled data but only a very limited amount of labeled training data. One general approach that has been explored for util...
Avrim Blum, John D. Lafferty, Mugizi Robert Rweban...
A method is given for constructing a max register, a linearizable, wait-free concurrent data structure that supports a write operation and a read operation that returns the larges...
Clustering methods can be either data-driven or need-driven. Data-driven methods intend to discover the true structure of the underlying data while need-driven methods aims at org...