Recent years has witnessed dramatic progress of neural machine translation (NMT), however, the method of manually guiding the translation procedure remains to be better explored. Previous works proposed to handle such problem through lexcially-constrained beam search in the decoding phase. Unfortunately, these lexically-constrained beam search methods suffer two fatal disadvantages: high computational complexity and hard beam search which generates unexpected translations. In this paper, we propose to learn the ability of lexically-constrained translation with external memory, which can overcome the above mentioned disadvantages. For the training process, automatically extracted phrase pairs are extracted from alignment and sentence parsing, then further be encoded into an external memory. This memory is then used to provide lexically-constrained information for training through a memory-attention machanism. Various experiments are conducted on WMT Chinese to English and English to German tasks. All the results can demonstrate the effectiveness of our method.
We propose an entanglement concentration protocol (ECP) for nonlocal N-electron systems in a partially entangled W state, resorting to ancillary single electrons and charge detection. Compared with other ECPs for W states, our ECP has some advantages. Firstly, only an entangled N-electron system is required in each round of entanglement concentration, not two copies. Secondly, only one of the users, say Charlie, needs to perform the protocol, while all parties should perform the same operations as Charlie in other ECPs for W-class states. Thirdly, only Charlie asks other parties to retain or discard their electrons, and they do not need to check their measurement results, which greatly simplifies the complication of classical communication. Fourthly, our ECP has a higher success probability than other ECPs for W-class states as the parties can recycle their concentration when they obtain a system in another W-class state with less entanglement. These advantages make our ECP more feasible for practical applications.
The image data transferred by embedded network devices directly lacks of extensibility. As different devices have different image formats, so the receivers may have difficulty in interpretation. This article proposes a method based on XML to solve this problem, researches coding and recoding of image and achieve image transferring by integrating the data into XML, finally validates the feasibility through an experiment.
A new filtering algorithm, adaptive square root cubature Kalman filter-Kalman filter (SRCKF-KF) is proposed to reduce the problems of amount of calculation, complex formula-transform, low accuracy, poor convergence or even divergence. The method uses cubature Kalman filter (CKF) to estimate the nonlinear states of model while its linear states are estimated by the Kalman filter (KF). The simulation and practical experiment results show that, compared to the extended Kalman filter (EKF) and unscented Kalman filter (UKF). The modified filter not only enhances the numerical stability, guarantees positive definiteness of the state covariance, but also increases accuracy, which has high practicability.
We propose to use a set of averaged entropies, the multiple entropy measures (MEMS), to partially quantify quantum entanglement of multipartite quantum state. The MEMS is vector-like with m = [N/2] components: [S1, S2, ..., Sm], and the i-th component Si is the geometric mean of i-qubits partial entropy of the system. The Si measures how strong an arbitrary i qubits from the system are correlated with the rest of the system. It satisfies the conditions for a good entanglement measure. We have analyzed the entanglement properties of the GHZ-state, the W-states, and cluster-states under MEMS.
The influence of the disturbance caused by the imperfection of the engineering coupling constants in the perfect state transfer is calculated. The results show that the fidelity for the perfect state transfer is seriously affected by the errors occurring near the input and output spins. Such results are helpful for the realization of the perfect state transfer in the case where there exist errors in experiments.
Due to the complexity of the brain image itself, the brain image segmentation technology has become a bottle for further application and development of the system. Considering the inconsistency of intensity, partial volume effect, and noise in medical images, this paper studies the brain image segmentation technology based on the multi-weight probability. The multi-weight probability method mainly models the data set with outliers and non-Gaussian noise. First, the probabilistic local ELM model is established. Based on this, the Parzen window method is used to establish the probability distribution of the local model, and then, the probability distribution is used as the weight to fuse. All local models are used to build a global robustness model. The method successfully applied the brain and UCI examples and compared with traditional ELM, regularized ELM, and robust ELM. The results show that the probability weight ELM shows better modeling performance.
In this paper the interacting multiple models algorithm and the converted measurement Kalman filter with debiasing are combined to realize the maneuvering target tracking with the non linear measuring equations. The measurements obtained in polar coordinates system is converted to rectangular coordinates system. The resulted data, with average error and variance being caculated out, is used for Kalman filtering after debiasing. The simulation proves that the mothod has a good result even if the measuring error were great.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)