摘要:
Speech recognition such as command and control speech recognition generally use a context free grammar to constrain the decoding process. Word or subword background model are constructed to repopulate dynamic hypothesis space, especially when word spareness is at issue. The background models can be later used in speech recognition. During speech recognition, background and conventional context free grammar decoding are used to measure confidence. The discussion above is merely provided for general background information and is not intended to be used as an aid in determining the scope of the claimed subject matter.
摘要:
A field emission cathode includes a conductive substrate and a carbon nanotube film disposed on a surface of the conductive substrate. The carbon nanotube film includes a plurality of successive and oriented carbon nanotube bundles parallel to the conductive substrate, the carbon nanotubes partially extrude from the carbon nanotube film. A method for fabricating the field emission cathode includes the steps of: (a) providing a conductive substrate; (b) providing at least one carbon nanotube film, the carbon nanotube film including a plurality of successive and oriented carbon nanotube bundles joined end to end, the carbon nanotube bundles parallel to the conductive substrate, and (c) disposing the at least one carbon nanotube film to the conductive substrate to achieve the field emission cathode.
摘要:
Multiple input modalities are selectively used by a user or process to prune a word graph. Pruning initiates rescoring in order to generate a new word graph with a revised best path.
摘要:
A surface-conduction electron emitter includes a substrate, two electrodes disposed on the substrate and parallel to each other, and a plurality of line-shaped carbon nanotube elements fixed on at least one electrode. One end of each carbon nanotube element points to the other electrode. An electron source using the surface-conduction electron emitter includes a substrate, a plurality of electrodes disposed on the substrate and parallel to each other, and a plurality of line-shaped carbon nanotube elements fixed on at least one electrode. One end of each carbon nanotube element points to the other electrode.
摘要:
There is provided a method and system for scheduling a job in a cluster, the cluster comprises multiple computing nodes, and the method comprises: defining rules for constructing virtual sub-clusters of the multiple computing nodes; constructing the multiple nodes in the cluster into multiple virtual sub-clusters based on the rules, wherein one computing node can only be included in one virtual sub-cluster; dispatching a received job to a selected virtual sub-cluster; and scheduling at least one computing node for the dispatched job in the selected virtual sub-cluster. Further, the job is dispatched to the selected virtual sub-cluster based on characteristics of the job and/or characteristics of virtual sub-clusters. The present invention can increase the throughput of scheduling effectively.
摘要:
A method of TI formation evaluation is disclosed. The method comprises receiving a plurality of borehole measurements; deriving a correlation between a first TI stiffness parameter and other TI stiffness parameters where the first and other TI stiffness parameters representing mechanical behavior of the TI formation; and computing the first and other TI stiffness parameters based on the borehole measurements and the derived correlation. The method further comprises evaluating TI formation elastic properties based on the computed first and other TI stiffness parameters. The method further comprises assuming that the shear modulus parallel to TI symmetric axis can be approximated from other moduli.
摘要:
A three-dimensional mask model of the invention provides a more realistic approximation of the three-dimensional effects of a photolithography mask with sub-wavelength features than a thin-mask model. In one embodiment, the three-dimensional mask model includes a set of filtering kernels in the spatial domain that are configured to be convolved with thin-mask transmission functions to produce a near-field image. In another embodiment, the three-dimensional mask model includes a set of correction factors in the frequency domain that are configured to be multiplied by the Fourier transform of thin-mask transmission functions to produce a near-field image.
摘要:
A solar collector includes a substrate having a top surface and a bottom surface opposite to the upper surface, a sidewall, a transparent cover, and a heat-absorbing layer. The sidewall is arranged on the top surface of the substrate. A transparent cover is disposed on the sidewall opposite to the substrate to form a sealed chamber with the substrate together. The heat-absorbing layer is disposed on the upper surface of the substrate and includes a carbon nanotube composite material.
摘要:
A multi-state pattern recognition model with non-uniform kernel allocation is formed by setting a number of states for a multi-state pattern recognition model and assigning different numbers of kernels to different states. The kernels are then trained using training data to form the multi-state pattern recognition model.
摘要:
A field emission device includes a transparent plate, an insulating substrate, one or more grids located on the insulating substrate. Each grid includes a first, second, third and fourth electrode down-leads and a pixel unit. The first, second, third and fourth electrode down-leads are located on the periphery of the grid. The first and the second electrode down-leads are parallel to each other. The third and the fourth electrode down-leads are parallel to each other. The pixel unit includes a phosphor layer, a first electrode, a second electrode and at least one electron emitter. The first electrode and the second electrode are separately located. The first electrode is electrically connected to the first electrode down-lead, and the second electrode is electrically connected to the third electrode down-lead. The phosphor layer is located on the corresponding first electrode.