Most up-to-date well-behaved topic-based summarization systems are built upon the extractive framework. They score the sentences based on the associated features by manually assig...
I present MOSES (meta-optimizing semantic evolutionary search), a new probabilistic modeling (estimation of distribution) approach to program evolution. Distributions are not esti...
Abstract. Inspired on psycholinguistics and neuroscience, a symbolicconnectionist hybrid system called θ-Pred (Thematic Predictor for natural language) is proposed, designed to re...
Numerous optimizations exist for improving the performance of nondeferred reference-counting (RC) garbage collection. Their designs are ad hoc, intended to exploit different count...
It is common for libraries to provide public access to historical and ancient document image collections. It is common for such document images to require specialized processing i...