This paper is devoted to the long time behavior for the Drift-diffusion semiconductor equations. It is proved that the dynamical system has a compact, connected and maximal attractor when the mobilities are constants ...This paper is devoted to the long time behavior for the Drift-diffusion semiconductor equations. It is proved that the dynamical system has a compact, connected and maximal attractor when the mobilities are constants and generation-recombination term is the Auger model; as well as the semigroup S(t) denned by the solutions map is differential. Moreover the upper bound of Hausdorff dimension for the attractor is given.展开更多
Drift phenomenon has been known as the drawback of sensors and causes inaccuracy on the long-term measurement. In general,there are two methods to reduce the drift problem.One is to tune the parameters of the fabricat...Drift phenomenon has been known as the drawback of sensors and causes inaccuracy on the long-term measurement. In general,there are two methods to reduce the drift problem.One is to tune the parameters of the fabrication process to improve the properties of the front-ended device.Another is to compensate the drift phenomenon by adding extra drift compensation circuit or software in the back-ended readout circuit.In this study,a drift calibration method for the potentiometric sensor was presented and the drift calibration method was performed by using the circuit.According to experimental results,the drift phenomenon of the SnO_2 pH electrode was reduced by the drift calibration device.展开更多
The purpose of tall Paper is to geueralize the Girsanov theorem to the case of a local martingale. The author uses the generalized Girsanov theorem to prove the existence ed uniqueness of diffusion processes with sigu...The purpose of tall Paper is to geueralize the Girsanov theorem to the case of a local martingale. The author uses the generalized Girsanov theorem to prove the existence ed uniqueness of diffusion processes with sigular drift terms.This result is much more general than the corresponding results in [4] and [8].展开更多
传统推荐算法多基于用户兴趣的静态属性获得用户偏好,忽略了用户兴趣漂移问题,为此,提出了解决该问题的融合用户兴趣漂移的Top-N推荐算法。利用长短期记忆网络(LSTM,long short term memory)处理时序数据的优势表示用户短期兴趣漂移规律...传统推荐算法多基于用户兴趣的静态属性获得用户偏好,忽略了用户兴趣漂移问题,为此,提出了解决该问题的融合用户兴趣漂移的Top-N推荐算法。利用长短期记忆网络(LSTM,long short term memory)处理时序数据的优势表示用户短期兴趣漂移规律,用矩阵分解得到的固定向量表示用户的长期兴趣,将注意力机制纳入LSTM隐藏状态的表示中来获取用户长短期兴趣关联。实验结果表明,所提算法与当前流行算法相比,在Top-N项目推荐中具有更优性能。展开更多
基金This work is supported by the Funds of the Nature Science Research of Henan(10371111).
文摘This paper is devoted to the long time behavior for the Drift-diffusion semiconductor equations. It is proved that the dynamical system has a compact, connected and maximal attractor when the mobilities are constants and generation-recombination term is the Auger model; as well as the semigroup S(t) denned by the solutions map is differential. Moreover the upper bound of Hausdorff dimension for the attractor is given.
文摘Drift phenomenon has been known as the drawback of sensors and causes inaccuracy on the long-term measurement. In general,there are two methods to reduce the drift problem.One is to tune the parameters of the fabrication process to improve the properties of the front-ended device.Another is to compensate the drift phenomenon by adding extra drift compensation circuit or software in the back-ended readout circuit.In this study,a drift calibration method for the potentiometric sensor was presented and the drift calibration method was performed by using the circuit.According to experimental results,the drift phenomenon of the SnO_2 pH electrode was reduced by the drift calibration device.
文摘The purpose of tall Paper is to geueralize the Girsanov theorem to the case of a local martingale. The author uses the generalized Girsanov theorem to prove the existence ed uniqueness of diffusion processes with sigular drift terms.This result is much more general than the corresponding results in [4] and [8].
文摘传统推荐算法多基于用户兴趣的静态属性获得用户偏好,忽略了用户兴趣漂移问题,为此,提出了解决该问题的融合用户兴趣漂移的Top-N推荐算法。利用长短期记忆网络(LSTM,long short term memory)处理时序数据的优势表示用户短期兴趣漂移规律,用矩阵分解得到的固定向量表示用户的长期兴趣,将注意力机制纳入LSTM隐藏状态的表示中来获取用户长短期兴趣关联。实验结果表明,所提算法与当前流行算法相比,在Top-N项目推荐中具有更优性能。
