We survey some recent results on efficient margin-based algorithms for adaptive sampling in binary classification tasks. Using the so-called Mammen-Tsybakov low noise condition to parametrize the distribution of covariates, and assuming linear label noise, we state bounds on the convergence rate of the adaptive sampler to the Bayes risk. These bounds show that, excluding logarithmic factors, th...

We present a novel scheme to combine neural machine translation (NMT) with traditional statistical machine translation (SMT). Our approach borrows ideas from linearised lattice minimum Bayes-risk decoding for SMT. The NMT score is combined with the Bayes-risk of the translation according the SMT lattice. This makes our approach much more flexible than n-best list or lattice rescoring as the neu...

We provide a new analysis of an efficient margin-based algorithm for selective sampling in classification problems. Using the so-called Tsybakov low noise condition to parametrize the instance distribution, we show bounds on the convergence rate to the Bayes risk of both the fully supervised and the selective sampling versions of the basic algorithm. Our analysis reveals that, excluding logarit...

Optimal risk and preventive patterns are itemsets which can identify characteristics of cohorts of individuals who have significantly disproportionate representation in the abnormal and normal groups. In this paper, we propose a new classifier namely ORPSW (Optimal Risk and Preventive Sets with Weights) to classify gene expression data based on optimal risk and preventive patterns. The proposed...

In this work, we propose a PAC-Bayes bound for the generalization risk of the Gibbs classifier in the multi-class classification framework. The novelty of our work is the critical use of the confusion matrix of a classifier as an error measure; this puts our contribution in the line of work aiming at dealing with performance measure that are richer than mere scalar criterion such as the misclas...

in this research, an iterative approach is employed to recognize and classify control chart patterns. to do this, by taking new observations on the quality characteristic under consideration, the maximum likelihood estimator of pattern parameters is first obtained and then the probability of each pattern is determined. then using bayes’ rule, probabilities are updated recursively. finally, when...

the objective of this study was to compare six statistical methods for prediction of genomic breedingvalues for traits with different genetic architecture in term of gene effects distributions and number ofquantitative traits loci (qtls). a genome consisted of 500 bi-allelic single nucleotide polymorphism(snp) markers distributed over a chromosomes with 100 cm length was simulated. three differ...