Fast machine-learning online optimization of ultra-cold-atom experiments.

We apply an online optimization process based on machine learning to the production of Bose-Einstein condensates (BEC). BEC is typically created with an exponential evaporation ramp that is optimal for ergodic dynamics with two-body s-wave interactions and no other loss rates, but likely sub-optimal for real experiments. Through repeated machine-controlled scientific experimentation and observations our 'learner' discovers an optimal evaporation ramp for BEC production. In contrast to previous work, our learner uses a Gaussian process to develop a statistical model of the relationship between the parameters it controls and the quality of the BEC produced. We demonstrate that the Gaussian process machine learner is able to discover a ramp that produces high quality BECs in 10 times fewer iterations than a previously used online optimization technique. Furthermore, we show the internal model developed can be used to determine which parameters are essential in BEC creation and which are unimportant, providing insight into the optimization process of the system.

Introduction of an embolus detection system based on analysis of the transcranial Doppler audio-signal.

A new embolus detection system (EDS) is presented, built with the intention of detecting ongoing cerebral embolization in patients at risk of transient ischaemic attacks or stroke. It is based on the analysis of the audio-Doppler signal of a transcranial Doppler machine. The algorithm of the EDS estimates the intensity, duration and zero-crossing dynamics of the audio signal. The EDS has a multi-layer neural network which classifies events into micro-emboli signals (MES) or artefacts. The decision-making component of the software has been validated against human experts. Data from patients in the post-operative phase of carotid surgery were used for the validation process. The results showed agreement in MES and artefact classification of > 93%. Apart from a monitoring display, the monitoring system includes a verification unit that allows the user to listen and to look at all data of individual MES and artefacts. Moreover, the system allows the user to record, store and re-calculate all data files. Data are stored using European Data Format, which allows data transportation over the Internet. The EDS may have a potential in stroke risk stratification, evaluating the effect of novel anti-thrombotic therapies, and in peri-operative and remote monitoring of carotid endarterectomy.