This paper presents a novel and notable swarm approach to evolve an optimal set of weights and architecture of a neural network for classification in data mining. In a distributed ...
Partitioning a memory into multiple blocks that can be independently accessed is a widely used technique to reduce its dynamic power. For embedded systems, its benefits can be ev...
Olga Golubeva, Mirko Loghi, Massimo Poncino, Enric...
The Fisher kernel is a generic framework which combines the benefits of generative and discriminative approaches to pattern classification. In this contribution, we propose to a...
Given a single program (i.e., bit string), one may assume that the program's behaviors can be determined by first identifying the native runtime architecture and then executi...
Sang Kil Cha, Brian Pak, David Brumley, Richard Ja...
Abstract— This paper describes experiments using reinforcement learning techniques to compute pattern urgencies used during simulations performed in a Monte-Carlo Go architecture...