Pawel Budzianowski

PyDial: A Multi-domain Statistical Dialogue System Toolkit

Statistical Spoken Dialogue Systems have been around for many years. However, access to these systems has always been difficult as there is still no publicly available end-to-end system implementation. To alleviate this, we present PyDial, an opensource end-to-end statistical spoken dialogue system toolkit which provides implementations of statistical approaches for all dialogue system modules. Moreover, it has been extended to provide multidomain conversational functionality. It offers easy configuration, easy extensibility, and domain-independent implementations of the respective dialogue system modules. The toolkit is available for download under the Apache 2.0 license.

Predictors of atrial fibrillation early recurrence following cryoballoon ablation of pulmonary veins using statistical assessment and machine learning algorithms

Inflammation, oxidative stress, myocardial injury biomarkers and clinical parameters (longer AF duration, left atrial enlargement, the metabolic syndrome) are factors commonly related to AF recurrence. This study aims to assess the predictive value of laboratory and clinical parameters responsible for early recurrence of atrial fibrillation (ERAF) following cryoballoon ablation (CBA) using statistical assessment and machine learning algorithms. This study group comprised 118 consecutive patients (mean age, 62.5u2009±u20097.8xa0years; women 36%) with paroxysmal (54.1%) and persistent (45.9%) AF who underwent their first pulmonary vein isolation (PVI) performed by CBA (Arctic Front Advance 2nd generation 28xa0mm). The biomarker concentrations were measured at baseline and after CBA in a 24-h follow-up. ERAF was defined as at least a 30-s episode of arrhythmia registered by a 24xa0h-Holter monitor within the 3xa0months following the procedure. 56 clinical, laboratory and procedural variables were collected from each patient. We used two classification algorithms: support vector machines, gradient boosted tree. The synthetic minority over-sampling technique (SMOTE) was used to provide a balanced training data set. Within a period of 3xa0months 21 patients (17.8%) experienced ERAF. The statistical analysis indicated that the lowered levels of post-ablation TnT (pu2009=u20090.043) and CK-MB (pu2009=u20090.010) with the TnT elevation (pu2009=u20090.044) were the predictors of ERAF following CBA. In addition, diabetes and statin treatment were significantly associated with ERAF after CBA (pu2009<u20090.05). The machine learning algorithms confirmed the results obtained in the univariate analysis.