Word and n-gram posterior probabilities estimated on N-best hypotheses have been used to improve the performance of statistical machine translation (SMT) in a rescoring framework....
Color is of interest to those working in computer vision largely because it is assumed to be helpful for recognition. This assumption has driven much work in color based image ind...
Efficiency is a prime concern in syntactic MT decoding, yet significant developments in statistical parsing with respect to asymptotic efficiency haven't yet been explored in...
In recent years, new semistatic word-based byte-oriented text compressors, such as Tagged Huffman and those based on Dense Codes, have shown that it is possible to perform fast d...
Abstract One of the main hurdles to improved CLIR effectiveness is resolving ambiguity associated with translation. Availability of resources is also a problem. First we present a ...