32results about How to "Probability is accurate" patented technology

A computer-implemented method for providing a candidate list of alternatives for a text selection containing text from multiple input sources, each of which can be stochastic (such as a speech recognition unit, handwriting recognition unit, or input method editor) or non-stochastic (such as a keyboard and mouse). A text component of the text selection may be the result of data processed through a series of stochastic input sources, such as speech input that is converted to text by a speech recognition unit before being used as input into an input method editor. To determine alternatives for the text selection, a stochastic input combiner parses the text selection into text components from different input sources. For each stochastic text component, the combiner retrieves a stochastic model containing alternatives for the text component. If the stochastic text component is the result of a series of stochastic input sources, the combiner derives a stochastic model that accurately reflects the probabilities of the results of the entire series. The combiner creates a list of alternatives for the text selection by combining the stochastic models retrieved. The combiner may revise the list of alternatives by applying natural language principles to the text selection as a whole. The list of alternatives for the text selection is then presented to the user. If the user chooses one of the alternatives, then the word processor replaces the text selection with the chosen candidate.

A method includes: obtaining process technology definition data related to a process technology of each layer forming a basic cell, from a process technology definition file defining process technology definition data related to a process technology for use in fabricating a semiconductor integrated circuit, thereby holding a process technology definition table; obtaining device structure data including data related to a device template which defines a structure of each layer of the basic cell and data related to the structure of the layer defined in accordance with the device template, from a device structure definition file, thereby holding the obtained device structure data as a device structure definition table; and determining the structure of each layer defined in accordance with the device template held as the obtained device structure data, thereby generating the layout pattern of the basic cell forming the semiconductor integrated circuit.

A base station creates and transmits power control data to a mobile terminal at each prescribed time slot based on signals received from the mobile terminal, and the mobile terminal controls the transmission power of signals to the base station based on the power control data. During this transmission power control, the mobile terminal measures the reception level of common pilot signals transmitted from the base station at each time slot, calculates a moving average of the reception level for a plurality of time slots, and when transmission to the base station is interrupted, stores the transmission power and moving average for the immediately preceding time slot, and when transmission to the base station is resumed, adds the difference between the moving average at the time of transmission resumption and the stored moving average to the stored transmission power, and uses the power obtained by this addition as the signal transmission power at the time of transmission resumption.

An encoding method and encoding apparatus for B-transform, and encoded data for same are provided. The apparatus is for encoding input data and includes: a B-transform unit for performing a B-transform on the input data to generate a binary expression; and an encoder for removing one or more binary symbols from the binary expression and scanning the binary expression with the one or more binary symbols removed to generate a bitstream. According to the method and apparatus, it is possible to encode uniform distribution data efficiently, and using the B-transform by its characteristics for reducing the bit number in the generation of the binary expressions from the B-transform operation on the uniform distribution data enhances the data compression performance.

An image processing apparatus comprises edge detection unit, on-axis chromatic aberration edge detection unit which detects an edge determined to include chromatic aberration on an axis from the edges, color deviation amount acquisition unit which acquires color deviation amounts on the edges, and correction unit which performs chromatic aberration correction, wherein the on-axis chromatic aberration edge detection unit detects, as the edge which includes the chromatic aberration on the axis, an edge having signals corresponding to at least one color on the edge of the image data, which has a blur amount that is not less than a threshold value, and signals corresponding to at least another color which has a blur amount that is less than the threshold value, and the color deviation amount acquisition unit acquires the color deviation amounts from edges among the edges.

A speech recognition and natural language understanding system performs insertion, deletion, and replacement edits of tokens at positions with low probabilities according to both a forward and a backward statistical language model (SLM) to produce rewritten token sequences. Multiple rewrites can be produced with scores depending on the probabilities of tokens according to the SLMs. The rewritten token sequences can be parsed according to natural language grammars to produce further weighted scores. Token sequences can be rewritten iteratively using a graph-based search algorithm to find the best rewrite. Mappings of input token sequences to rewritten token sequences can be stored in a cache, and searching for a best rewrite can be bypassed by using cached rewrites when present. Analysis of various initial token sequences that produce the same new rewritten token sequence can be useful to improve natural language grammars.

The invention provides a game log based user gender determination method and system. Different from an existing method, the game log based user gender determination method is characterized in that game gender polarity is introduced, and user gender is speculated through a weighting method; the game gender polarity can be directly acquired by making statistics of a specific name of a user use game, and user gender accuracy measured according to the game gender polarity is quite high; model mining performed on existing user information construction data is not needed, special training on the model is omitted, the gender of a corresponding user can be accurately acquired through simple statistics and calculation, use of complex models is avoided, and excessive system cost is avoided; with the method, the user can be accurately subjected to user group positioning with low cost by a game operator through log information, the game is pushed in a targeted manner, and running cost is greatly reduced; meanwhile, user experience can be enhanced.