Effects of sublingual glyceryl trinitrate (GTN) were studied in ten patients without heart failure during diagnostic cardiac catheterisation following angiography. GTN caused substantial reduction in peak left ventricular and aortic pressure (19 mmHg) with lesser reduction in mean aortic pressure (9 mmHg) and no change in diastolic aortic pressure. Reduction in stroke volume (by 15%), associated with fall in left ventricular end diastolic pressure (by 4 mmHg) was insufficient to explain the marked (17 mmHg - 34%) reduction in pulse pressure. Decrease in pulse pressure was associated with loss of the late systolic peak on both the aortic and left ventricular pressure wave. This peak is caused by pulse wave reflection. GTN caused no change in peripheral resistance or in indices of aortic compliance (characteristic impedance, total arterial compliance) but was associated with reduction in fluctuations of both modulus and phase of aortic impedance. All these changes in pressure waves and in impedance spectra are explicable on the basis of decreased peripheral wave reflection. This can be attributed to the known vasodilatory effect of GTN on the peripheral arteries. Simulation of arterial vasodilatation in a multi-branched model of the systemic arterial system confirmed this interpretation. Dilatation of peripheral arteries explains in part the beneficial effects of GTN in adult man.

We introduce a reinforcement learningbased approach to simultaneous machine translation-producing a translation while receiving input wordsbetween languages with drastically different word orders: from verb-final languages (e.g., German) to verb-medial languages (English). In traditional machine translation, a translator must "wait" for source material to appear before translation begins. We remove this bottleneck by predicting the final verb in advance. We use reinforcement learning to learn when to trust predictions about unseen, future portions of the sentence. We also introduce an evaluation metric to measure expeditiousness and quality. We show that our new translation model outperforms batch and monotone translation strategies.

No abstract