We propose a nonparallel data-driven emotional speech conversion method. It enables the transfer of emotion-related characteristics of a speech signal while preserving the speaker's identity and linguistic content. Most existing approaches require parallel data and time alignment, which is not available in many real applications. We achieve nonparallel training based on an unsupervised style transfer technique, which learns a translation model between two distributions instead of a deterministic one-to-one mapping between paired examples. The conversion model consists of an encoder and a decoder for each emotion domain. We assume that the speech signal can be decomposed into an emotion-invariant content code and an emotion-related style code in latent space. Emotion conversion is performed by extracting and recombining the content code of the source speech and the style code of the target emotion. We tested our method on a nonparallel corpora with four emotions. The evaluation results show the effectiveness of our approach.

No abstract

The wind speed forecast is the basis of the wind power forecast. The wind speed has the characteristics of random non-smooth so obviously that its precise forecast is extremely difficult. Therefore, a forecasting method based on the theory of chaotic phase-space reconstruction and SVM was put forward in this paper and a forecasting model of Chaotic Support Vector Machine was built. In order to improve the precision and generalization ability, the key parameters in the phase space reconstruction and the key parameters of SVM were carried out joint optimization by using particle swarm algorithm in the paper. Then the optimal parameters were brought into the forecasting model to forecast short-term wind speed. The above method was applied to wind speed forecast of a wind farm in Inner Mongolia, China. In the experiments of computer simulation, the absolute percentage error of forecasting results was only 12.51%, which showed this method was effective for short-term wind speed forecast.

With the rapid growth of cloud email services, email encryption is beginning to be used more and more to alleviate concerns about cloud privacy and security. However, this increase in usage invites the problem of how to search and filter encrypted emails effectively. Searchable public key encryption is a popular technology to solve encrypted email searching, but encrypted email filtering is still an open problem. We propose an encrypted cloud email searching and filtering scheme based on hidden policy ciphertextpolicy attribute-based encryption with keyword search as a new solution. It enables the recipient to search the encrypted cloud email keywords and allows the email filtering server to filter the encrypted email content when receiving the email, as the traditional email keyword filtering service. Our hidden policy scheme is constructed by composite order bilinear groups and proven secure by dual system encryption methodology. Our scheme can be applied to other scenarios such as file searching and filtering and has certain practical value.INDEX TERMS attribute-based keyword search; Dual System Encryption; encrypted email filtering; hidden policy.

No abstract