Self-supervised approaches for speech representation learning are challenged by three unique problems: (1) there multiple sound units in each input utterance, (2) is no lexicon of during the pre-training phase, and (3) have variable lengths with explicit segmentation. To deal these problems, we propose Hidden-Unit BERT (HuBERT) approach self-supervised learning, which utilizes an offline cluste...

We propose a generative model approach to contour tracking against non-stationary clutter and to coping with occlusions by explicit modelling and inferring. The proposed dynamic Bayesian networks consist of multiple hidden processes which model the target, the clutter and the occlusions. The image observation models, which depict the generation of the image features, are conditioned on all the ...

When communities engage in risk-sharing with asymmetric information, wherein a member of a risk-sharing group is unable to verify the income and shocks experienced by the others, they are unable to achieve full smoothing of income risk (Thomas and Worrall 1990). The frictions that drive the possible negative impacts of hidden income are likely to be mediated by the relationships among a risk-sh...

The mobile nodes are infrequent movement in nature; therefore, its packet transmission is also infrequent. Packet overload occurred for routing process, and data lossed by receiver node, since hackers hide the normal node. Basically, hidden node problem created based on malicious that planned to vital relay specific path. loss routing; so, it minimizes delivery ratio network lifetime. Then, pro...

To explain conceptual gap between classical/quantum and other, hypothetical descriptions of world, several principles has been proposed. So far, all these principles have not involved uncertainty concept. Here we introduce an information content principle (ICP) which represents the new-constrained uncertainty principle. The principle, by taking into account the encoding/decoding properties of p...