635 results about "Adaptation method" patented technology

The invention provides a mobile terminal and a method for self-adaptively increasing the download rate thereof. The mobile terminal includes a network monitoring module, a rate detection module and a link switching arbitration unit; the network detection module detects wireless network signals around the mobile terminal; the rate detection module detects the network rate; the link switching The arbitrating unit switches the link for the mobile communication terminal to provide wireless data services to the best link according to the above detection results. The technical solution of the present invention does not require additional protocol development on mobile devices, saves development time, and has better versatility and standardization, and can be widely used in mobile devices with universal access functions, such as mobile phones, data Network cards and other products.

The invention relates to a gateway of the Internet of things, which comprises a standard interface connector for inserting wireless sensing units, an information receiving module for receiving information of the wireless sensing units through the standard interface connector, a protocol query model for querying applicable protocol stacks from a management platform of the Internet of things at the network side according to the received information of the wireless sensing units, a protocol deployment module for downloading the queried protocol stacks applicable to the wireless sensing units into the gateway of the Internet of things for deployment, and a protocol conversion and adaptation module for carrying out protocol conversion and adaptation between nodes in a wireless sensor network and the management platform of the Internet of things according to the deployed protocol stacks and a message protocol format defined by the management platform of the Internet of things. In the invention, the gateway of the Internet of things can insert different wireless sensing units according to the specific need of the wireless sensor network and queries and downloads the protocol stacks corresponding to the wireless sensing units from the management platform of the Internet of things so as to adapt to various wireless networks rapidly.

An adaptation method includes using a first classifier trained on projected representations of labeled objects from a first domain to predict pseudo-labels for unlabeled objects in a second domain, based on their projected representations. A classifier ensemble is iteratively learned. The ensemble includes a weighted combination of the first classifier and a second classifier. This includes training the second classifier on the original representations of the unlabeled objects for which a confidence for respective pseudo-labels exceeds a threshold. A classifier ensemble is constructed as a weighted combination of the first classifier and the second classifier. Pseudo-labels are predicted for the remaining original representations of the unlabeled objects with the classifier ensemble and weights of the first and second classifiers in the classifier ensemble are adjusted. As the iterations proceed, the unlabeled objects progressively receive pseudo-labels which can be used for retraining the second classifier.

A speech recognition method and apparatus perform speaker clustering and speaker adaptation using average model variation information over speakers while analyzing the quantity variation amount and the directional variation amount. In the speaker clustering method, a speaker group model variation is generated based on the model variation between a speaker-independent model and a training speaker ML model. In the speaker adaptation method, the model in which the model variation between a test speaker ML model and a speaker group ML model to which the test speaker belongs which is most similar to a training speaker group model variation is found, and speaker adaptation is performed on the found model. Herein, the model variation in the speaker clustering and the speaker adaptation are calculated while analyzing both the quantity variation amount and the directional variation amount. The present invention may be applied to any speaker adaptation algorithm of MLLR and MAP.

A fiber optic system includes a transmitter configured to utilize a plurality of modulation formats and a receiver communicatively coupled to the transmitter and configured to utilize a plurality of modulation formats. The transmitter and the receiver are cooperatively configured to set a modulation format of the plurality of modulation formats based upon optical signal-to-noise ratio associated therewith. A flexible bandwidth adaptation method includes monitoring at least one aspect of an optical link at a network element, responsive to the at least one aspect, computing a new modulation scheme for the optical link, and, if a solution is found for the new modulation scheme, changing to the new modulation format.

An object of the present invention is to facilitate dealing with noisy speech with varying SNR and save calculation costs by generating a speech model with a single-tree-structure and using the model for speech recognition.Every piece of noise data stored in a noise database is used under every SNR condition to calculate the distance between all noise models with the SNR conditions and the noise-added speech is clustered. Based on the result of the clustering, a single-tree-structure model space into which the noise and SNR are integrated is generated (steps S1 to S5). At a noise extraction step (step S6), inputted noisy speech to be recognized is analyzed to extract a feature parameter string and the likelihoods of HMMs are compared one another to select an optimum model from the tree-structure noisy speech model space (step S7). Linear transformation is applied to the selected noisy speech model space so that the likelihood is maximized (step S8).

A speech recognition adaptation method for a vehicle having a telematics unit with an embedded speech recognition system. Speech is received and pre-processed to generate acoustic feature vectors, and an adaptation parameter is applied to the acoustic feature vectors to yield transformed acoustic feature vectors. The transformed acoustic feature vectors are decoded and a hypothesis of the speech is selected, and the adaptation parameter is trained using acoustic feature vectors from the hypothesis. The method also includes one or more of the following steps: the speech is observed for a certain characteristic and the trained adaptation parameter is saved in accordance with the certain characteristic for use in transforming feature vectors of subsequent speech having the certain characteristic; use of the trained adaptation parameter persists from one vehicle ignition cycle to the next; and use of the trained adaptation parameter is ceased upon detection of a system fault.

The invention discloses an Internet of Things adaptation method and system, computer equipment and a storage medium. The method comprises the following steps: establishing a plurality of protocol adaptation plug-ins of the Internet of Things, wherein the protocol adaptation plug-ins are used for carrying out format conversion on Internet of Things data and server data; when target data transmission between the Internet of Things and the server is carried out, determining a target protocol adaptation plug-in of the target Internet according to the target Internet where the target data is located; performing format conversion on the target data according to the target protocol adaptation plug-in; and sending the data after format conversion to a destination address. The protocol adaptation plug-ins can be established for each Internet of Things, data interaction between one server and multiple devices in the Internet of Things can be achieved by establishing the multiple adaptation plug-ins, and linkage of the devices in the Internet of Things of different manufacturers is achieved.

An apparatus and method for eliminating the spurious signal of a non-linear amplifier in a base station transmitter of a mobile communication system are provided. A digital predistorter includes first and second predistorters, each having a memoryless non-linearity part using a look-up table and a memory non-linearity part using an finite impulse response (FIR) filter. The first predistorter adaptively updates the look-up tables, while the second predistorter adaptively updates the filtering coefficients of the FIR filters. According to another embodiment of the present invention, a digital predistorter has a memoryless non-linearity part using a look-up table and a memory non-linearity part using an FIR filter, and adaptively updates the look-up table and the filtering coefficients of the FIR filter.

The invention belongs to the field of machine vision and intelligent control for achieving fast self-adaptation of traffic video monitoring target detection. The method comprises the steps of building an initial training sample bank; training an AdaBoost classifier based on Haar characteristics and a support vector machine (SVM) classifier based on histograms of oriented gradients (HOG) characteristics respectively; and detecting monitoring images frame by frame by employing of the two classifiers, wherein the detection process is divided into steps of predicting of sub-images in a detection frame through the two classifiers respectively, performing of confidence-degree determination on predicted results, adding of prediction labels corresponding to large confidence-degree and the sub-images into additional training sample banks of the classifiers corresponding to small confidence-degree till the size of the detection frame reaches half the sizes of detected images, retraining of the two classifiers by using the updated training sample banks and detecting of a next frame of image till all images are detected, and the final classifiers can be used for detecting targets of vehicles, pedestrians and the like in actual traffic scenes.

Disclosed is a link margin adaptation method using feedback information of a wireless communication system. For transmission and reception of data between two nodes of the wireless communication system, a transmission node requests a receiving node to transmit a link margin. The transmission node receives the link margin, and selects one of a white noise table or a delay spread table according to a delay spread value between the two nodes. Then, the transmission node adapts a link between the two nodes to a transmission mode and decides a transmission rate to transmit data at the decided transmission rate, thereby improving transmission capacity of the wireless communication system, obtaining the optimum link state, extending power utilization time of stations, and reducing interference between stations.

The example embodiment of the present invention provides an acoustic model adaptation method for enhancing recognition performance for a non-native speaker's speech. In order to adapt acoustic models, first, pronunciation variations are examined by analyzing a non-native speaker's speech. Thereafter, based on variation pronunciation of a non-native speaker's speech, acoustic models are adapted in a state-tying step during a training process of acoustic models. When the present invention for adapting acoustic models and a conventional acoustic model adaptation scheme are combined, more-enhanced recognition performance can be obtained. The example embodiment of the present invention enhances recognition performance for a non-native speaker's speech while reducing the degradation of recognition performance for a native speaker's speech.

New layer of data link layer protocol - data link procedure (DLP) is introduced in the invention to realize syncretizing adaptation between Ethernet and SDH / SONET in order to overcome shortage of current adaptation technique including not matched interface speed, not supporting dynamic bandwidth allocation based on packets, and not possible to transmit SDH / SONET on Ethernet, not compatible to IP and packet voice service etc. The method realizes direct transmission of Ethernet on SDH / SONET, and SDH / SONET on Ethernet as well as realizes compatibility to IP network and packet voice network. Using safety mechanism, network control management mechanism, flow management mechanism provided by DLP realizes network management control function such as two layer protected rearrangement, performance management, fault management etc. The method makes current communication network transits to united public network in next generation of using packet switching technology smoothly.

The invention provides an unsupervised field adaptation method and system based on adversarial learning and a medium, and the method comprises the steps: a feature extraction step: employing a featureextraction network to extract the features of images in a source field and a target field, and obtaining the image features of the source field and the image features of the target field; a categoryprediction step: predicting the probability that the image belongs to each category according to the obtained image feature of the source field and the image feature of the target field, and obtaininga category prediction probability; and a field discrimination step: predicting the probability that the image feature comes from the source field and the target field through a field discrimination network according to the obtained image feature of the source field and the image feature of the target field, and obtaining a field prediction probability. According to the method, the characteristicsthat the image extraction fields in the source field and the target field are invariable and have relatively high discrimination capability can be realized, so that the unsupervised field adaptationis realized.

The invention relates to a camera and robot control system based on a semantic model and an automatic adaptation method thereof. According to the method, a semantic model server-side module and a semantic model client-side module are built in a robot controller and a camera controller respectively, and the automatic adaptation of the industrial camera and the robot control system is realized through semantic modeling of the semantic model server-side module and the analysis of the semantic model client-side module. Under the condition that the function of the camera is upgraded and changed, the robot controller does not need to stop working to carry out programming on the controller again to modify the function and the command of a camera, the working flexibility of a robot system is improved, the working time is saved, and plug-and-play is achieved.

The invention relates to a UA adaptation method of a web page. The method comprises the steps as follows: step A, a network server obtains UA information or information identification of a mobile phone which sends out an access request; step B, the network server searches and obtains the characteristic information of the mobile phone according to the UA information or information identification; step C, the network server carries out page adaptation according to the characteristic of the mobile phone; and step D, the network server returns the adaptive page to the mobile phone, so as to adjust page width, image size, image quality, paragraph parameter, character color, typeface, size, audio and / or video according to different parameter and / or attribute settings and ensure that the page is adapted to the corresponding mobile phone. The invention is adapted to different mobile phone characteristics and ensures that the user can obtain the web page adapted to own mobile phone.

The invention discloses a mobile terminal webpage adaptation system and a mobile terminal webpage adaptation method. The mobile terminal webpage adaptation method comprises the following steps of: performing preset adaptation on webpage of a URL (uniform resource locator) corresponding to a server side, and generating adaptive webpage; browsing a webpage request transmitted from a mobile terminal to the server side and monitored by a controller to obtain request information in the webpage request, and requesting HTML (Hyper Text Mark-up Language) tag data from the server side; returning the HTML tag data of webpage contents to a browser manager controller after the server side obtains the HTML tag data of the webpage contents; and browsing the HTML tag data of the webpage contents returned from the server side and captured by the controller, converting the HTML tag data of the webpage contents, transmitting the converted data to a webpage adapter to be subjected to adaption processing, returning the HTML tag data subjected to the conversion and adaption processing to the mobile terminal, analyzing the HTML tag data by virtue of a browser on the mobile terminal, and then rendering and displaying the HTML tag data on the mobile terminal. The mobile terminal webpage adaptation system and the mobile terminal webpage adaptation method can be used for conveniently and quickly realizing the webpage adaptation of the mobile terminal.

An unsupervised adaptation method and apparatus are provided that reduce the storage and time requirements associated with adaptation. Under the invention, utterances are converted into feature vectors, which are decoded to produce a transcript and alignment unit boundaries for the utterance. Individual alignment units and the feature vectors associated with those alignment units are then provided to an alignment function, which aligns the feature vectors with the states of each alignment unit. Because the alignment is performed within alignment unit boundaries, fewer feature vectors are used and the time for alignment is reduced. After alignment, the feature vector dimensions aligned to a state are added to dimension sums that are kept for that state. After all the states in an utterance have had their sums updated, the speech signal and the alignment units are deleted. Once sufficient frames of data have been received to perform adaptive training, the acoustic model is adapted.

The invention relates to an adaptation method for network television services, which comprises: receiving an access request that a set-top box accesses IPTV service; initiating a capability inquiry request of inquiring the set-top box capacity information, and initiating a content inquiry request of inquiring the IPTV content requested for using in access request; initiating a template inquiry request of inquiring a display template that matches with the set-top box capacity information according to the returned capacity information of the set-top box; and adapting the IPTV content according to the returned display template and issuing the adapted IPTV content to the set-top box. The invention also relates to a network television service adapter server and a system. The adaptable display template can be selected according to the set-top box capacity information, the IPTV content requested by the set-top box can be adapted by the display template, so the service system can provide services in uniform way to the set-top box with different software and hardware capacities and reduce development costs of services and difficulty of running and maintenance of the set-top box.

According to one aspect of the invention, a method is provided in which a set of multiple mixture monophone models is created and trained to generate a set of multiple mixture context dependent models. A set of single mixture triphone models is created and trained to generate a set of context dependent models. Corresponding states of the triphone models are clustered to obtain a set of tied states based on a decision tree clustering process. Parameters of the context dependent models are estimated using a data dependent maximum a posteriori (MAP) adaptation method in which parameters of the tied states of the context dependent models are derived by adapting corresponding parameters of the context independent models using the training data associated with the respective tied states.

An object of the present invention is to facilitate dealing with noisy speech with varying SNR and save calculation costs by generating a speech model with a single-tree-structure and using the model for speech recognition. Every piece of noise data stored in a noise database is used under every SNR condition to calculate the distance between all noise models with the SNR conditions and the noise-added speech is clustered. Based on the result of the clustering, a single-tree-structure model space into which the noise and SNR are integrated is generated (steps S1 to S5). At a noise extraction step (step S6), inputted noisy speech to be recognized is analyzed to extract a feature parameter string and the likelihoods of HMMs are compared one another to select an optimum model from the tree-structure noisy speech model space (step S7). Linear transformation is applied to the selected noisy speech model space so that the likelihood is maximized (step S8).

The invention provides an updated system for mobile phone network game platform and a method thereof, both of which are applied in the development of the mobile online game and provide the unified user login, user registry, game hall, friend-making and chatting, network communication, game match, game currency mode and mobile terminal adaptation method and greatly shortens the time and reduces the difficulty in adaptability development of the mobile phone game on the mobile terminals of different brands and different types by extracting the generality of the mobile online game and encapsulating the difference of the mobile terminals. The system also provides an innovative system aiming at the currency of mobile phone games and anti-fraud practice, which ensures that the playability of the mobile online game is stronger and the competitiveness more outstanding. Therefore, the invention solves a series of problems that the development cost is high, the development cycle is long, the stability is not high, the adaptability is bad, the user operation is not unified, the experience property is not good, etc. when the mobile phone network game is developed and provides a platform-integrated solution to the mobile network game.

A speech recognition adaptation method for a vehicle having a telematics unit with an embedded speech recognition system. Speech is received and pre-processed to generate acoustic feature vectors, and an adaptation parameter is applied to the acoustic feature vectors to yield transformed acoustic feature vectors. The transformed acoustic feature vectors are decoded and a hypothesis of the speech is selected, and the adaptation parameter is trained using acoustic feature vectors from the hypothesis. The method also includes one or more of the following steps: the speech is observed for a certain characteristic and the trained adaptation parameter is saved in accordance with the certain characteristic for use in transforming feature vectors of subsequent speech having the certain characteristic; use of the trained adaptation parameter persists from one vehicle ignition cycle to the next; and use of the trained adaptation parameter is ceased upon detection of a system fault.

The invention relates to a pressure adaptation method for an electric heating cooker and a device thereof. In the method, a boiling point control parameter is set in advance; a main control chip reads the conversion of the boiling point control parameter and carries out cooking according to the converted corresponding liquid boiling temperature; the device comprises a hardware switch which is connected with the main control chip and is used for setting the corresponding boiling point control parameter or a memory which is connected with or arranged in the main control chip and is used for storing the boiling point control parameter. The method and the device thereof set the boiling point control parameter of the electric heating cooker according to different atmospheric pressures (altitude). The main control chip recognizes the parameter and carries out the corresponding heating adjustment. Therefore, the cooker can work normally under various atmospheric pressures (altitude).

The invention provides an open set domain adaptation method based on entropy minimization. The open set domain adaptation method comprises the steps of creating an open set domain adaptation network, and initializing network hyper-parameters; inputting a source domain picture and a target domain picture into an open set domain adaptive network, obtaining picture depth features through a networkfeature extraction layer, and then obtaining prediction probability vectors of various categories through a softmax layer according to a feature map; calculating an entropy loss function by using a known category prediction probability vector and a picture label so as to correctly identify a common category; on the basis, calculating binary cross entropy and a binary diversity loss function by using an unknown category prediction probability value so as to correctly classify unknown categories, updating network parameters by using a back propagation gradient until a loss function value is minimum, and stopping training the network; and storing the network model and the training result, and introducing the target domain data set into the network model to obtain a final target domain label.

The invention discloses a scenario adaptation device and a scenario adaptation method for a user interface. The device comprises a facial feature identification module, a scenario customization module, a feature-scenario matching module and a scenario application module, wherein the facial feature identification module is used for identifying facial features of a user, and transmitting identified user facial feature information to the feature-scenario matching module; the scenario customization module is used for recording mapping relationships between user facial feature information and customized scenario information, and different user facial features correspond to different customized scenarios; the feature-scenario matching module is used for receiving the user facial feature information from the facial feature identification module, matching the user facial feature information and the user facial feature information in each mapping relationship in the scenario customization module one by one, and transmitting the matched customized scenario information to the scenario application module if the user facial feature information is matched with the user facial feature information in a certain mapping relationship; the scenario application module is used for applying a scenario on the user interface according to the customized scenario information when the customized scenario information is received from the feature-scenario matching module. According to the device and the method, the user interface of mobile equipment is adaptively switched.

A speech recognition method and apparatus perform speaker clustering and speaker adaptation using average model variation information over speakers while analyzing the quantity variation amount and the directional variation amount. In the speaker clustering method, a speaker group model variation is generated based on the model variation between a speaker-independent model and a training speaker ML model. In the speaker adaptation method, the model in which the model variation between a test speaker ML model and a speaker group ML model to which the test speaker belongs which is most similar to a training speaker group model variation is found, and speaker adaptation is performed on the found model. Herein, the model variation in the speaker clustering and the speaker adaptation are calculated while analyzing both the quantity variation amount and the directional variation amount. The present invention may be applied to any speaker adaptation algorithm of MLLR and MAP.

Methods and systems for processing signals in a receiver are disclosed herein and may include generating a plurality of chip-rate synchronously sampled signals utilizing a plurality of received clusters. At least a portion of the generated plurality of chip-rate synchronously sampled signals may be simultaneously equalized in time domain and in spatial domain. The equalization may be based on a plurality of weight values calculated for the plurality of received clusters. The weight values may be iteratively computed utilizing a time-based adaptation method, such as a conjugate gradient (CG) search. The equalized portion of the generated plurality of chip-rate synchronously sampled signals may be added to generate a total equalized signal. The total equalized signal may be demodulated to generate a demodulated signal. At least one convolutional code and / or turbo code within the demodulated signal may be decoded.

The invention discloses an interface adaptation method, a device, a computer device and a storage medium on IOS device. The method comprises the following steps of: obtaining the resolution of the IOSdevice to be identified, including a first screen height and a first screen width, obtaining the resolution of the UI interface to be adapted and the reference IOS device configured on the referenceIOS device, wherein the resolution of the reference IOS device comprises a second screen height and a second screen width, acquiring a height adaptation ratio based on the first screen height and thesecond screen height, and acquiring a width adaptation ratio based on the first screen width and the second screen width; acquiring a ratio difference value of the height adaptation ratio and the width adaptation ratio, adjusting the interface height and the interface width of the UI interface to be adapted according to the judgment result of the ratio difference value and the preset difference value, and acquiring a target UI interface; on the IOS device to be identified, the target UI interface is adapted to be displayed. The method is adapted according to the adaptation ratio, and has better adaptation effect.

The invention discloses a feature and instance joint transfer learning method in a semi-supervised scene. The method aims at solving the problem of insufficient labeled data during classification model training in some fields. Data in other fields and label-free data in the field are introduced to assist training. Meanwhile, the difference of data distribution between the fields is considered, andfor the situation that a small amount of data with labels and a large amount of label-free data exist in a target domain, the invention provides a mixed balanced distribution adaptation method and aself-learning instance migration method. A feature and instance joint transfer learning method FSJT is constructed on the basis.