Abstract: One or more processors receive recording data of a meeting between a first individual and a second individual. One or more processors analyze the recording data to identify one or more characteristics of the first individual. One or more processors match the one or more characteristics of the first individual to one or more preferences of a third individual.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for performing dynamic, stroke-based alignment of touch displays. In one aspect, a method includes obtaining a candidate transcription that an automated speech recognizer generates for an utterance, determining a particular context associated with the utterance, determining that a particular n-gram that is included in the candidate transcription is included among a set of undesirable n-grams that is associated with the context, adjusting a speech recognition confidence score associated with the transcription based on determining that the particular n-gram that is included in the candidate transcription is included among the set of undesirable n-grams that is associated with the context, and determining whether to provide the candidate transcription for output based at least on the adjusted speech recognition confidence score.

Abstract: A system and method for generating a transcript of an audio input. An embodiment of a system and method may include generating a phonetic lattice by decoding the audio input and producing a transcription based on the phonetic lattice and based on a first language model. A transcription may be analyzed to produce analysis results. Analysis results may be used to select from a plurality of language models, one language model and the selected language model may be used to generate a transcript of the audio input.

Abstract: Provided is a lightweight computational device that is configured to be in communication with a cloud both directly and via a smart computational device. The lightweight computational device receives a voice command from a user, wherein the lightweight computational device does not have adequate processing power to convert the voice command to a text command. The voice command is transmitted from the lightweight computational device to a smart computational device, wherein the smart computational device uses voice recognition to convert the voice command to a text command in the smart computational device, and transmits the text command for being processed by that cloud that provides at least one of voice recognition service and other services. The lightweight computational device receives a data response for the user from the cloud, via the smart computational device, based on the other services provided by the cloud.

Abstract: A computer-implemented method for enrolling in a database voice prints generated from audio streams may include receiving an audio stream of a communication session and creating a preliminary association between the audio stream and an identity of a customer that has engaged in the communication session based on identification information. The method may further include determining a confidence level of the preliminary association based on authentication information related to the customer and if the confidence level is higher than a threshold, sending a request to compare the audio stream to a database of voice prints of known fraudsters. If the audio stream does not match any known fraudsters, sending a request to generate from the audio stream a current voice print associated with the customer and enrolling the voice print in a customer voice print database.

Abstract: Various systems, apparatuses, and methods for training classifiers using selected cohort sample subsets are disclosed herein, in an example, a set of target supervectors, representing a target class, is received, and a set of cohort supervectors, representing a cohort class, is received. A distance metric is calculated from a respective cohort supervector to a respective target supervector, and a proper subset of cohort supervectors are selected based on the calculated distance metrics. The set of target supervectors and the selected proper subset of cohort supervectors are used to train a classifier. Further examples described herein describe how training classifiers using selected cohort sample subsets may be used to increase performance and decrease resource consumption in voice biometric systems.

Abstract: The present document discloses a method and apparatus for compensating for a lost frame in a transform domain, comprising: calculating frequency-domain coefficients of a current lost frame using frequency-domain coefficients of one or more frames prior to the current lost frame, and performing frequency-time transform to obtain an initially compensated signal; and performing waveform adjustment, to obtain a compensated signal. Alternatively, extrapolation is performed for all or part of frequency points of the current lost frame using phases and amplitudes of corresponding frequency points of a plurality of previous frames to obtain phases and amplitudes of the corresponding frequency points of the current lost frame, to obtain frequency-domain coefficients of the corresponding frequency points, and frequency-time transform is performed to obtain a compensated signal.

Abstract: A linear prediction coefficient of a signal represented in a frequency domain is obtained by performing linear prediction analysis in a frequency direction by using a covariance method or an autocorrelation method. After the filter strength of the obtained linear prediction coefficient is adjusted, filtering may be performed in the frequency direction on the signal by using the adjusted coefficient, whereby the temporal envelope of the signal is shaped. This reduces the occurrence of pre-echo and post-echo and improves the subjective quality of the decoded signal, without significantly increasing the bit rate in a bandwidth extension technique in the frequency domain represented by SBR.

Abstract: Increasing the level of the consonant segments relative to the nearby vowel segments, known as consonant-vowel ratio (CVR) modification, is reported to be effective in improving speech intelligibility by listeners in noisy backgrounds and by hearing-impaired listeners. A method along with a system for real-time CVR modification using the rate of change of spectral centroid for detection of spectral transitions is disclosed. A preferred embodiment of the invention using a 16-bit fixed point processor with on-chip FFT hardware is also presented for real-time signal processing. It can be integrated with other FFT-based signal processing in communication devices, hearing aids, and other systems for improving speech perception under adverse listening conditions.

Abstract: A signal processing apparatus for dereverberating a number of input audio signals, where the signal processing apparatus includes a processor configured to transform the number of input audio signals into a transformed domain to obtain input transformed coefficients, the input transformed coefficients being arranged to form an input transformed coefficient matrix, determine filter coefficients upon the basis of eigenvalues of a signal space, the filter coefficients being arranged to form a filter coefficient matrix, convolve input transformed coefficients of the input transformed coefficient matrix by filter coefficients of the filter coefficient matrix to obtain output transformed coefficients, and the output transformed coefficients being arranged to form an output transformed coefficient matrix.

Abstract: A method includes identifying an event generated by a software application executed by a processor in a device. The event is not associated with a user interaction with the device. Without requesting or receiving user input from a user of the device, a microphone of the device is enabled for a predetermined time period after identifying the occurrence of the event to identify a user voice command. The user voice command is executed. A method includes generating an event alert notification on a speaker of a device. Without querying a user of the device, a microphone of the device is enabled for a predetermined time period after generating the event alert notification to identify a user voice command. The user voice command is executed.

Abstract: A process for slicing a row from an array of devices includes positioning an array such that a row physically interfaces with a conforming fence, applying a force to the fence to conform it to the mating face of the row, applying a vacuum force to the fence to secure it in conformal position with the row, and then slicing the row from the array. Applying the force to the fence to conform it to the row inhibits leakage associated with the vacuum force utilized to secure the fence with the row. A stronger hold of the row is provided, which can lead to more precise slicing of the row. A slicing tool includes a rotatable support that, in operation, is supported by an air bearing and hence is able to freely rotate to a position such that the fence conforms to a workpiece face.

Abstract: A device including a magnetic structure, the magnetic structure having a substrate adjacent surface and a second, opposing surface, the magnetic structure having a near field transducer (NFT), wherein the NFT includes gold or an alloy thereof, and is positioned at the second surface an overcoat structure; and a film structure, the film structure positioned between the magnetic structure and the overcoat structure, the film structure having a total thickness of not greater than about 100 ?, and the film structure including: a first interfacial structure having a first and a second opposing surface; a second interfacial structure having a first and a second opposing surface; and an intermediate structure wherein the first surface of the first interfacial structure is positioned adjacent the NFT of the magnetic structure, and the second surface of the second interfacial structure is positioned adjacent the overcoat structure, and the intermediate structure is positioned between the first interfacial structure an

Abstract: Implementations described and claimed herein include a reader structure, comprising a first reader, including a sensor stack and a top shield structure, the top shield structure comprises a synthetic antiferromagnetic shield (SAF) structure, including a reference layer including at least a layer of NiFe and an impurity additive, an RKKY coupling layer RKKY coupling layer (e.g., Ru layer), and a pinned layer. In another implementation, the RL of the SAF shield structure of a first reader includes at least a layer of amorphous magnetic material. Yet, in another implementation, the SAF shield structure includes an insertion layer of amorphous magnetic material under the SAF shield RL, within the SAF shield RL or between the SAF shield RL and SAF shield Ru.

Abstract: A hard disk drive actuator pivot-base tower clearance spacer mechanism may be positioned at least in part between an enclosure base tower structure, on which a pivot shaft of a pivot bearing assembly is disposed, and the pivot shaft. The spacer mechanism is positioned to affect the clearance between an actuator of which the pivot bearing assembly is part and the base tower structure, such as to limit unwanted tilting of the actuator relative to the tower. The clearance spacer mechanism may comprise an elastic cap or may comprise a spring mechanism configured to fit over the top of the base tower structure so as to compress the spacer mechanism while positioned between the tower structure and the pivot shaft.

Abstract: According to one embodiment, an adjustment method is applied to a magnetic disk device includes a disk, and a head including a heater and configured to protrude to the disk based on amount of heat the heater varies. The method includes calculating a first adjust amount by subtracting a difference, for spacing the head from the disk, from a first control amount applied to the heater when the head contacts a first area, calculating a second adjust amount by subtracting the difference from a second control amount applied to the heater when the head contacts a second area, and adjusting a time when the second adjust amount is applied to the heater, based on a result of comparison between the first and second control amounts.

Abstract: In recording technologies for batch formation of a plurality of recording bits in a recording medium by forming a plurality of optical spots using an ultrashort pulse laser and a spatial optical modulator, the hatch recordable hit number has an upper limit, resulting in restricted recording speed. The intensity of the optical spot is corrected to increase the batch recordable bit number for increasing the recording speed.

Abstract: Various devices and systems may benefit from enhanced reading of optical media. For example, certain computer systems may benefit from array reading of optical media. An apparatus may include, for example, an array of optical sensors. The array of optical sensors may be configured to read a plurality of parallel linear strips of data from an optical medium.

Abstract: Embodiments generally relate to data storage in a computing system. The present technology discloses techniques that that can enable an optimized mechanism to change spinning speed of data storage disk drives. The present technology can use a service controller, e.g. a Baseboard Management Device (BMC), to communicate with a disk controller to change the spinning speed of disk drives. The present technology can improve energy efficiency by efficiently controlling the spinning speed of disk drives. It can also reduce data access latency by promptly spinning up a disk from a spun-down state.

Abstract: An apparatus includes a loop filter that receives a plurality of input signals. Each of the input signals is based on a different timing error detector output signal. The apparatus also includes a plurality of read channels, a plurality of interpolation filters, and an array of transducers. Each of the interpolation filters is in communication with a corresponding one of the read channels. Each of the transducers is in communication with a corresponding one of the read channels. The loop filter processes the plurality of input signals, and outputs a different total phase signal for each received input signal. Each of the interpolation filters samples the corresponding read channel based on one of the total phase signals output by the loop filter. The loop filter processes the plurality of input signals by calculating a phase estimate of the samples, and a skew estimate of the samples, relative to written data.

Abstract: To read data from a storage medium, a magnetization map of the data stored on at least one portion of the storage medium is obtained. The magnetization map is analyzed to obtain intermediate data corresponding to the data stored on the at least one portion of storage medium. Further, the intermediate data is converted into the user data based on format emulation of a data storage format of the storage medium. The conversion of the intermediate data includes decrypting the intermediate data to obtain the user data.

Abstract: A method of decoding a codeword that satisfies a k constraint into output data includes, using a decoder of a hard disk drive system, converting each bit of the codeword in Non-Return-to-Zero format, extracting, from the converted codeword, a plurality of data blocks comprising a first data block and a plurality of remaining data blocks, performing a first analysis on the plurality of data blocks for modifying each of the plurality of data blocks that satisfies a first predetermined criterion, and performing a second analysis on the plurality of data blocks after the first analysis for modifying each of the plurality of data blocks that satisfies a second predetermined criterion to obtain the output data.

Abstract: A playback device includes a playback controller that sets a scene with a higher level of importance at a lower playback speed and sets a scene with a lower level of importance at a higher playback speed based on a level of importance of each scene of video data, and an image data processor that converts the video data into a playback speed set by the playback controller for each scene, wherein the playback controller plays back the video data within a preset playback time. Further, the playback method plays back video data within a preset playback time by playing back a scene with a higher level of importance at a lower playback speed and playing back a scene with a lower level of importance at a higher playback speed based on a level of importance of each scene.

Abstract: A video editing system and method using machine learning are described. In one embodiment, the video editing system comprises editing processing logic controllable to perform at least one edit on one or more raw input feeds to render one or more final cut clips for viewing, each edit to transform data from one or more of the raw input feeds into the one or more of the plurality of final cut clips by generating tags that identify highlights from signals and a machine learning module operable to access data from memory and operable to generate settings to control the editing processing logic based on the data using one or more machine learning algorithms to control the editing processing logic.

Abstract: A video editing apparatus and method using time-based highlight identification are described. In one embodiment, a video editing apparatus comprises a memory to store first video data, a time mapper module operable to determine first time information associated with the first video data, the first time information specifying a time frame during which the video data was captured, a communication interface to receive highlight list data corresponding to the time frame, an extractor operable to extract media clip data from the first video data based on the highlight list data, and a composer operable to compose the movie with the media clip data.

Abstract: A Video Editing Apparatus with Participant Sharing is described. In one embodiment, the video editing system comprises a memory to store instructions and video data that captures an activity of a participant and one or more processing units coupled to the memory, the one or more processing units executing the instructions to determine existence of one or both of signal data and media of a co-participant in an activity and to implement editing processing logic to create a clip from the video data by processing signals and editing the video data, where the processing of the signals and the editing of the video data are based on one or more of signal data and media associated with the co-participant in the activity.

Abstract: Methods for processing captured video data based on capture device orientation are described. In one embodiment, the method comprises capturing video data with a video capture device, detecting orientation of the video capture device, mapping pixels of the video data captured to a landscape orientation if the video capture device is in a portrait orientation, and displaying the video data on a screen of the video capture device in landscape orientation regardless of the orientation of the video capture device.

Abstract: A video editing system and method using multi-stakeholder control are described. In one embodiment, the video editing system comprises editing processing logic controllable to perform, on one or more raw input feeds, a plurality of different edits to render one or more final cut clips for viewing, where each of the one or more edits transforms data from one or more of the raw input feeds into the one or more of the plurality of final cut clips by generating tags that identify highlights from signals, and generating one or more variations of the final cut clips as a result of independent control and application of the editing processing logic to data from the one or more raw input feeds, wherein the highlights are generated based on a master highlight list generated based on processing of tags from the tagging.

Abstract: A video editing system and method using multi-stakeholder, multi-stage control are described. In one embodiment, the video editing system comprises editing processing logic controllable to perform, on one or more raw input feeds, a plurality of different edits to render one or more final cut clips for viewing, where the editing processing logic is part of each of a plurality of stages of an editing process that is responsive to a plurality of stakeholders interacting with the tagging and the highlights to generate the plurality of final cut streams, and each of the one or more edits to transform data from one or more of the raw input feeds into the one or more of the plurality of final cut clips by generating tags that identify highlights from signals, and generating one or more variations of the final cut clips as a result of independent control and application of the editing processing logic to data from the one or more raw input feeds.

Abstract: A video editing apparatus and method for generating multiple final cut clips is described. In one embodiment, a video editing system comprises editing processing logic controllable to perform, on one or more raw input feeds, a plurality of different edits to render one or more final cut clips for viewing, where each of the one or more edits transforms data from one or more of the raw input feeds into the one or more of the plurality of final cut clips by generating tags that identify highlights from signals, and generating one or more variations of the final cut clips as a result of independent control and application of the editing processing logic to data from the one or more raw input feeds.

Abstract: In some embodiments, a video processing system may be configured to process a video of a sporting event or match to determine a start point and an end point of each of a plurality of events within the match. The video processing system may automatically determine data associated with each event of the plurality of events based on the video and store the data in a memory. In some embodiments, the video processing system may segment the video into a plurality of video clips such that each video clip corresponds to one of the events of the plurality of events.

Abstract: A video editing system and method with multi-stage control to generate clips are described.

Abstract: Disclosed are a method and electronic device for video definition switching. The method comprises: loading a first fragment file of a first definition video from a pre-set second level cache region to a video stream cache region to be played; receiving a definition switching instruction; acquiring a configuration file of a second definition video from a video server according to the instruction; determining a second fragment file which corresponds to the first fragment file and meets a smooth switching condition in the second definition video; sequentially loading all fragment files arranged after the second fragment file in the second definition video to the second level cache region; and loading the fragment files to the video stream cache region after the first fragment file is played completely, and starting to play the second definition video.

Abstract: A video editing method using participant sharing is described. In one embodiment, the method comprises determining existence of one or both of signal data and media of a co-participant in an activity, obtaining video data that captures an activity of a participant, and creating a clip from the video data by processing signals and editing the video data, where the processing of the signals and the editing of the video data are based on one or more of signal data and media associated with the co-participant in the activity.

Abstract: A recording device includes an imaging unit that takes a video, a recording time conversion processor that converts a frame rate of specified video data in the video so as to adjust the video to fall within a specified recording time, and a tag range processor that sets a tag range where a tag can be assigned in accordance with a level of importance to the video data, and the recording time conversion processor converts a frame rate at a first thinning-out rate that thins out a frame in a first video data range where the tag range is not set in the video data, and converts a frame rate at a second thinning-out rate that thins out a frame in a second video data range where the tag range is set in the video data in accordance with the level of importance.

Abstract: Embodiments are directed towards presenting a current-playing-position marker on a progress bar along with real time content. An initial timeline value of the content may be determined, along with an initial presentation timestamp value for the first frame of the content. For each next frame of the content, a next timeline value may be determined based on a difference between a next presentation timestamp value for a corresponding next frame and the initial presentation timestamp value, and mapping of the next presentation timestamp value for the corresponding next frame and the next timeline value in the index file. The content and a progress bar may be displayed to the user on a remote device based on the index file. A position of the current-playing-position marker on the progress bar for each frame of the content may be determined based on the next timeline value for each corresponding frame.

Abstract: There are provided a memory system and an operating method thereof. A memory system includes a memory device suitable for storing data therein; and a memory controller suitable for initializing the memory device, or maintaining or changing a mode of the memory device according to power of the memory device during a wake-up operation.

Abstract: Systems and methods are described for compensating for variations in process, voltage, temperature, or combinations thereof in an apparatus. An example apparatus may be a memory circuit. A pre-driver circuit and driver circuit may be associated with the memory circuit. A reference generator may provide the pre-driver circuit with reference signals that are insensitive to process, voltage, and temperature. The pre-driver circuit may receive the reference signals and the pre-driver circuit output ramping rate may then be made less sensitive to variations in process, voltage, and temperature. The pre-driver circuit output may then be supplied to a driver circuit that may then output a final driver data output with reduced noise.

Abstract: Examples of the present disclosure provide apparatuses and methods for simulating access lines in a memory. An example method can include receiving a first bit-vector and a second bit-vector in a format associated with storing the first bit-vector in memory cells coupled to a first access line and a first number of sense lines and storing the second bit-vector in memory cells coupled to a second access line and the first number of sense lines. The method can include storing the first bit-vector in a number of memory cells coupled to the first access line and a second number of sense lines and storing the second bit-vector in a number of memory cells coupled to the first access line and a third number of sense lines, wherein a quantity of the first number of sense lines is less than a quantity of the second and third number of sense lines.

Abstract: A current sense amplifier is provided comprising a reference current input terminal, a control line input terminal, a sense current input terminal and a first output terminal. The amplifier further comprises a first NAND gate comprising first and second gate input terminals, and a second output terminal being coupled to the first output terminal of the amplifier. The amplifier also comprises two cross coupled inverters each comprising an n-FET, an n-FET input terminal, and each n-FET having a respective source. The amplifier further comprises a transmission gate comprising two transmission terminals and a gate terminal. The gate terminal is coupled to the control line terminal.

Abstract: A memory device includes: a plurality of first lines; a plurality of second lines; a plurality of bank groups each including a predetermined number of banks; and a column signal transmission unit suitable for transmitting one or more column command signals and one or more column address signals to the bank groups through the first lines based on an odd-numbered column command, and transmitting the column command signals and the column address signals to the bank groups through the second lines based on an even-numbered column command.

Abstract: A magnetoresistive memory device includes a variable resistance element and a read circuit. The resistance element has a resistance state, which is one of switchable first and second resistance states. The first and second resistance states exhibit different resistances. Each of the first and second resistance states is reached by a current flowing through the variable resistance element in one of opposing first and second directions. The read circuit passes a read current through the variable resistance element autonomously in the first or second direction in accordance with the resistance state of the variable resistance element.

Abstract: A non-volatile memory device according to an embodiment includes a first conductive layer, a second conductive layer including metal nitride, the metal nitride absorbing oxygen, a paraelectric layer disposed between the first conductive layer and the second conductive layer, a ferroelectric layer disposed between the paraelectric layer and the second conductive layer, the ferroelectric layer including hafnium oxide, at least one third conductive layer disposed on opposite side of at least one of the first conductive layer and the second conductive layer to the ferroelectric layer, the at least one third conductive layer including metal oxide, the metal oxide having oxygen ratio larger than stoichiometric ratio, and a sense circuit configured to read data based on tunneling current flow between the first conductive layer and the second conductive layer through the paraelectric layer and the ferroelectric layer.

Abstract: A memory system is disclosed, which relates to a technology for reducing current consumption needed to perform a refresh operation in a Dual In-line Memory Module. The memory system includes a memory module, which includes a volatile memory and a non-volatile memory, and a host configured to provide a refresh command to the memory module during the refresh operation. The memory module is configured to store cell characteristic information of the volatile memory in the non-volatile memory in an idle state, and control a refresh operation of the volatile memory in response to the cell characteristic information.

Abstract: Circuits and methods are described for a DDR memory controller where two different DQS gating modes are utilized. These gating modes together ensure that the DQS signal, driven by a DDR memory to the memory controller, is only available when read data is valid. Two types of gating logic are used: Initial DQS gating logic, and Functional DQS gating logic. The Initial gating logic has additional timing margin in the Initial DQS gating value to allow for the unknown round trip timing during initial bit levelling calibration. DQS functional gating is then optimized during further calibration to gate DQS precisely as latency and phase calibration are performed, resulting in a precise gating value for Functional DQS gating. Providing dual gating modes is especially useful when data capture is performed at half the DQS frequency in view of rising clock rates for DDR memories.

Abstract: A stacked semiconductor device is disclosed that includes a plurality of semiconductor dies. Each die has oppositely disposed first and second surfaces, with pads formed on each of the surfaces. A plurality of through-vias connect respective pads on the first surface to respective pads on the second surface. The through-vias include a first group of through-vias coupled to respective I/O circuitry on the semiconductor die and a second group of through-vias not coupled to I/O circuitry on the semiconductor die. The plurality of semiconductor dies are stacked such that the first group of through-vias in a first one of the plurality of semiconductor dies are aligned with respective ones of at least a portion of the second group of through-vias in a second one of the plurality of semiconductor dies.

Abstract: The present disclosure is directed to apparatuses and methods to pre-compensate and/or post-compensate for inter-symbol interference (ISI). The apparatuses and methods of the present disclosure can be implemented as part of an I/O interface in a memory or memory controller, including dynamic random access memory (DRAM).

Abstract: Methods of managing data of a storage device responsive to temperature can include measuring a temperature of the storage device, changing a duration of a refresh interval of the buffer memory responsive to the measured temperatures, changing a number of refresh bursts during the refresh interval responsive to the measured temperature, and refreshing data of the buffer memory based on the refresh interval and the number of the refresh bursts that are changed responsive to temperature.

Abstract: Approaches for stability of cells in a Static Random Access Memory (SRAM) array are provided. A circuit includes a precharging circuit configured to precharge bitlines of a Static Random Access Memory (SRAM) array to a first voltage potential for a non-read operation and to a second voltage potential for a read operation. The first voltage potential is different than the second voltage potential.

Abstract: A semiconductor device includes a memory cell; circuitry that generates a reference voltage; and a sense amplifier including a first input terminal electrically connected to the memory cell, and a second input terminal electrically connected to the circuitry. The sense amplifier obtains a value in correlation with a resistance value of the memory cell based on a comparison between a sense voltage applied to the first input terminal and the reference voltage applied to the second input terminal. The sense voltage changes at a speed in correlation with the resistance value of the memory cell. In at least part of a period during which the sense voltage changes, the circuitry causes the reference voltage to change in a direction opposite to a direction in which the sense voltage changes.