of twelve feature selection methods (e.g. Information Gain) evaluated on a benchmark of 229 text classification problem instances that were gathered from Reuters, TREC, OHSUMED, etc. The results are analyzed from multiple goal perspectives—accuracy, F-measure, precision, and recall—since each is appropriate

Snakes, or active contours, are used extensively in computer vision and image processing applications, particularly to locate object boundaries. Problems associated with initialization and poor convergence to boundary concavities, however, have limited their utility. This paper presents a new

to the correct marginals. However, on the QMR network, the loopy be liefs oscillated and had no obvious relation ship to the correct posteriors. We present some initial investigations into the cause of these oscillations, and show that some sim ple methods of preventing them lead to the wrong results

In contrast to expectation-based, predictive views of discourse comprehension, a model is developed in which the initial processing is strictly bottom-up. Word meanings are activated, propositions are formed, and inferences and elaborations are produced without regard to the discourse context

-linearities allowing them to compactly represent highly non-linear and highly-varying functions. However, until recently it was not clear how to train such deep networks, since gradient-based optimization starting from random initialization appears to often get stuck in poor solutions. Hin-ton et al. recently

Syntactic natural language parsers have shown themselves to be inadequate for processing highly-ambiguous large-vocabulary text, as is evidenced by their poor per- formance on domains like the Wall Street Journal, and by the movement away from parsing-based approaches to textprocessing

Persistent microvascular hyperpermeability to plasma proteins even after the cessation of injury is a characteristic but poorly understood feature of normal wound healing. It results in extravasation of fibrinogen that clots to form fibrin, which serves as a provisional matrix and promotes

exceed the current state of the art retrieval performance on these datasets, particularly on queries with poor initial recall where techniques like query expansion suffer. Overall we show that soft-assignment is always beneficial for retrieval with large vocabularies, at a cost of increased storage

collection, useless occur-check elimination), program transformation (e.g., partial evaluation, parallelization), andeven program correctness proofs (e.g., termination proof). After a few simple introductory examples, we recall the classical framework for abstract interpretation of programs. Starting from a

poor health have significantly lower educational attainment, poorer health, and lower social status as adults. Childhood health and circumstance appear to operate both through their impact on initial adult health and economic status, and through a continuing direct effect of prenatal and childhood