He, WA
Dandan He, Redmond, WA US
Patent application number | Description | Published |
---|---|---|
20110258575 | DISPLAYING LARGE DATASETS IN A BROWSER ENVIRONMENT - A mechanism for displaying large datasets may download an approximate range of the dataset along with a subset of the dataset to a client. The client may display the subset as well as indicators for the approximate range. When a user moves to a position within the dataset that is not downloaded, the client may send a request for a second subset of the dataset, which may be displayed on the client. The user may interact with the dataset to select different items in the dataset, for example, and those interactions may be collected while the user views different portions of the dataset. The user may process all of the interactions over the entire dataset or a portion of the dataset by committing the changes. | 10-20-2011 |
20130159380 | PROVISIONING HIGH PERFORMANCE COMPUTING CLUSTERS - Provisioning a cloud based high performance computing cluster. The method includes from a cloud based provisioning service deployed in a cloud based computing provider, providing a user interface. The method further includes receiving user input at the cloud based provisioning service from the user interface. The method further includes from the user input, determining a configuration for a cloud based user service for the user. Using the cloud based provisioning service, the method further includes provisioning the cloud based user service for the user, in the cloud, based on the determined configuration. | 06-20-2013 |
20140207859 | PROVISIONING HIGH PERFORMANCE COMPUTING CLUSTERS - Provisioning a cloud based high performance computing cluster. The method includes from a cloud based provisioning service deployed in a cloud based computing provider, providing a user interface. The method further includes receiving user input at the cloud based provisioning service from the user interface. The method further includes from the user input, determining a configuration for a cloud based user service for the user. Using the cloud based provisioning service, the method further includes provisioning the cloud based user service for the user, in the cloud, based on the determined configuration. | 07-24-2014 |
Dingshan He, Bellevue, WA US
Patent application number | Description | Published |
---|---|---|
20080256427 | SYSTEM, METHOD, AND SERVICE FOR PROVIDING A GENERIC RAID ENGINE AND OPTIMIZER - A generic RAID engine system accepts an access request, accepts a metadata input comprising a layout description and, optionally, a plurality of resource optimization objectives, accepts a dynamic input comprising a dynamic state of an I/O stack comprising the generic RAID engine and a fault configuration of a plurality of storage devices in the I/O stack, and accepts RAID code input comprising information about the RAID code used by the I/O stack. The metadata input, the dynamic input, and the RAID code input are utilized to transform the access request into individual device reads and individual device writes such that RAID code relationships for the storage devices are maintained at all times. An optional optimizer module selects strategies that meet the resource optimization objectives. | 10-16-2008 |
20080270704 | CACHE ARRANGEMENT FOR IMPROVING RAID I/O OPERATIONS - The embodiments of the invention provide a method, apparatus, etc. for a cache arrangement for improving RAID I/O operations. More specifically, a method begins by partitioning a data object into a plurality of data blocks and creating one or more parity data blocks from the data object. Next, the data blocks and the parity data blocks are stored within storage nodes. Following this, the method caches data blocks within a partitioned cache, wherein the partitioned cache includes a plurality of cache partitions. The cache partitions are located within the storage nodes, wherein each cache partition is smaller than the data object. Moreover, the caching within the partitioned cache only caches data blocks in parity storage nodes, wherein the parity storage nodes comprise a parity storage field. Thus, caching within the partitioned cache avoids caching data blocks within storage nodes lacking the parity storage field. | 10-30-2008 |
20080270878 | CACHE ARRANGEMENT FOR IMPROVING RAID I/O OPERATIONS - The embodiments of the invention provide a method, apparatus, etc. for a cache arrangement for improving RAID I/O operations. More specifically, a method begins by partitioning a data object into a plurality of data blocks and creating one or more parity data blocks from the data object. Next, the data blocks and the parity data blocks are stored within storage nodes. Following this, the method caches data blocks within a partitioned cache, wherein the partitioned cache includes a plurality of cache partitions. The cache partitions are located within the storage nodes, wherein each cache partition is smaller than the data object. Moreover, the caching within the partitioned cache only caches data blocks in parity storage nodes, wherein the parity storage nodes comprise a parity storage field. Thus, caching within the partitioned cache avoids caching data blocks within storage nodes lacking the parity storage field. | 10-30-2008 |
Jerry He, Redmond, WA US
Patent application number | Description | Published |
---|---|---|
20110202509 | EFFICIENT EXTRACTION AND COMPRESSION OF DATA - A device for dynamically extracting and compressing information for a streaming media asset is provided. One embodiment of the device provides a computing device comprising a processor and memory comprising instructions stored therein that are executable by the processor. The instructions stored in the memory are executable to provide to a requesting computing device dynamically compressed information for a streaming media asset, the dynamically compressed information derived from an information file comprising variable data elements arranged in one or more data fields according to a well-known structure. For example, the instructions are executable to receive from the requesting computing device a request for the compressed information, extract the variable data elements from the information file, compress the variable data elements to form compressed data elements, and send to the requesting computing device a compressed file comprising the compressed data elements. | 08-18-2011 |
Jialin He, Redmond, WA US
Patent application number | Description | Published |
---|---|---|
20100240413 | Smart Card File System - An application programming interface (API) may receive high level file commands and implement those commands using the storage mechanism on a smart card. The smart card may have a processor and storage mechanism and may communicate to a host device using a packet based communication protocol, such as ADPU. The API may translate the high level file commands into one or more ADPU commands, communicate with the smart card, receive APDU responses, and translate the responses into high level file commands. A high level file command may allow access to a file using long file names, a hierarchical directory structure, and may allow creating, writing, reading, and deleting a file. Some embodiments may have more complex functions for navigating and manipulating a hierarchical directory structure, as well as defining metadata including access privileges and file types to individual files. | 09-23-2010 |
Jianmin He, Redmond, WA US
Patent application number | Description | Published |
---|---|---|
20100211204 | METHOD AND SYSTEM FOR ROBOTIC ASSEMBLY PARAMETER OPTIMIZATION - A method and system to optimize the parameters of a robot used in an assembly process. The assembly process is categorized based on its nature which may be cylindrical, radial and multi-stage insertion. The search pattern and search parameters are specified. The parameters are optimized and the optimized parameter set are verified and when a predetermined criteria such as assembly cycle time set and/or success rate is met the optimization process stops. When the optimization stops the verified parameters are used to cause the robot to perform the categorized assembly process. If the parameters do not meet the predetermined criteria, another round of optimization using the same or other parameters can be performed. | 08-19-2010 |
Jianmin He, Bellevue, WA US
Patent application number | Description | Published |
---|---|---|
20100145509 | Method And System For In-Production Optimization of The Parameters Of A Robot Used for Assembly - A robot is used to repeatedly assemble part during a continuous production run of parts to be assembled. There are parameters of the robot associated with the assembly. These parameters are used to assemble the parts. Simultaneously with that repeated assembly the robot parameters are optimized. The parts to be assembled have a starting position for the assembly and the simultaneous optimization of the robot assembly parameters also includes the learning of the starting position. The robot assembly parameters can be optimized in a predefined manner. The optimized parameters can then be verified and the optimized and verified robot assembly parameters may then be used in place of the parameters of the robot associated with the parts assembly along with the learned starting position. | 06-10-2010 |
20150217452 | METHOD AND SYSTEM FOR ROBOTIC ASSEMBLY PARAMETER OPTIMIZATION - A method and system to optimize the parameters of a robot used in an assembly process. The assembly process is categorized based on its nature which may be cylindrical, radial and multi-stage insertion. The search pattern and search parameters are specified. The parameters are optimized and the optimized parameter set are verified and when a predetermined criteria such as assembly cycle time set and/or success rate is met the optimization process stops. When the optimization stops the verified parameters are used to cause the robot to perform the categorized assembly process. If the parameters do not meet the predetermined criteria, another round of optimization using the same or other parameters can be performed. | 08-06-2015 |
20150217453 | METHOD AND SYSTEM FOR ROBOTIC ASSEMBLY PARAMETER OPTIMIZATION - A method and system to optimize the parameters of a robot used in an assembly process. The assembly process is categorized based on its nature which may be cylindrical, radial and multi-stage insertion. The search pattern and search parameters are specified. The parameters are optimized and the optimized parameter set are verified and when a predetermined criteria such as assembly cycle time set and/or success rate is met the optimization process stops. When the optimization stops the verified parameters are used to cause the robot to perform the categorized assembly process. If the parameters do not meet the predetermined criteria, another round of optimization using the same or other parameters can be performed. | 08-06-2015 |
Liwei He, Bellevue, WA US
Patent application number | Description | Published |
---|---|---|
20120150854 | Relevance Estimation using a Search Satisfaction Metric - The subject disclosure is directed towards using a satisfaction model's prediction as to whether a user was satisfied or dissatisfied in satisfying a search goal to help estimate the relevance of a URL/document that was returned and clicked by the user. The clickthrough data for a search goal session is processed by either a utility model or a despair model based on whether the satisfaction model indicated that the search goal session ended with the user satisfied or dissatisfied, respectively. The utility model distributes a utility value to each clicked URL, while the despair model distributes a despair value to each clicked URL. The utility value and despair value of each query-URL pair may be used as corresponding feature data for learning a search ranker. | 06-14-2012 |
20130173573 | Search Engine Performance Evaluation Using a Task-based Assessment Metric - The subject disclosure is directed towards evaluating search engine implementation performance using a task-based assessment metric. Logged user activities associated with the search engine implementation are processed into sessions and atomic events corresponding to the user activities within the sessions. The atomic events corresponding to the user activities are classified into tasks based on similarity of the queries within the user activities. After applying the task-based assessment metric to task information associated with the search engine implementation, an evaluation mechanism determines indicia of search engine implementation performance in terms of user behavior. | 07-04-2013 |
Quan He, Lacey, WA US
Patent application number | Description | Published |
---|---|---|
20110209839 | Method for making absorbent products - A method of making personal hygiene articles of high absorbency from cellulosic fiber pulp without chemical cross-linking of the cellulosic fibers is disclosed. The method comprises interposing a sublayer of mechanically fluffed alkali extracted cellulosic fiber pulp between a fluid permeable layer and a fluid impermeable layer. | 09-01-2011 |
Rongsen He, Pullman, WA US
Patent application number | Description | Published |
---|---|---|
20090141711 | Interleaved Multistage Switching Fabrics And Associated Methods - An interleaved multistage switching fabric includes Y multistage switching fabric panels, where Y is an integer greater than one. Each panel has primary inputs for receiving cells to be routed, local outputs for outputting routed cells, primary outputs for outputting non-routed cells, and reentry points for introducing non-routed cells into the panel. The switching fabric additionally includes at least one demultiplexer subsystem communicatively coupled to primary inputs of each panel, for interfacing the switching fabric with input lines. The switching fabric further includes at least one multiplexer subsystem communicatively coupled to local outputs of each panel, for interfacing the switching fabric with destination queues. The switching fabric additionally includes Y recirculation connections, where each recirculation connection communicatively couples primary outputs of one panel to reentry points of another panel. | 06-04-2009 |
Shaojian He, Bellevue, WA US
Patent application number | Description | Published |
---|---|---|
20150262256 | COLLABORATIVE INSIGHTS FOR ADVERTISEMENT VIEWERS - Systems, methods, and computer-readable storage media for managing advertisement viewer feedback are provided. The system includes a database and a server. The database stores advertisement data, including metrics, for each advertisement. The servers provide advertisements to an advertisement viewer. In turn, the server receives actions from the advertisement viewer in response to the advertisement. The metrics for the advertisement are provided for display based on actions received from the advertisement viewer. The advertisement is rendered such that upon user interaction the metrics become viewable by the user. Theses metrics may include collaborative insights that are measured from one or more other users with similar interests to the user viewing the selected advertisement. | 09-17-2015 |
Tony He, Kirkland, WA US
Patent application number | Description | Published |
---|---|---|
20150017962 | INTELLIGENT SIM SELECTION SUPPORTING RICH CONTEXT OF INPUT FACTORS - Various user interfaces and other technologies for interacting with devices that support multiple SIM cards can be implemented. For example, intelligent handling of SIM card selection can help users deal with multiple SIM card scenarios. A user interface for overriding SIM card selection can be supported. Users can interact with their devices more efficiently and with less frustration. A wide variety of use default SIM card scenarios are supported. SIM cards can be selected based on any of a rich set of input factors. | 01-15-2015 |
20150018040 | INTELLIGENT USER INTERFACES FOR MULTIPLE SIM CARDS - Various user interfaces and other technologies for interacting with devices that support multiple SIM cards can be implemented. For example, SIM-separated and SIM-aggregated user interface paradigms can be supported. Intelligent selection of an appropriate paradigm can support user preferences, conversation user interfaces, and the like. Other features such as SIM card defaults can help users deal with multiple SIM card scenarios. A consistent, compact user interface for switching SIM cards can be supported. Users can interact with their devices more efficiently and with less frustration. A wide variety of use scenarios are supported. | 01-15-2015 |
20150049158 | SEAMLESS CALL TRANSITIONS - Various user interfaces and other technologies for seamlessly transitioning between calls of different types can be implemented. The technologies can be implemented to give the impression of a single call that is upgraded from one call type to another. A new application can register so that an appropriate user interface control appears for activation when seamless call transition is possible. Transitioning for third-party applications can thus be supported. Cross-platform implementations can be supported. | 02-19-2015 |
20150049867 | SEAMLESS CALL TRANSITIONS - Various user interfaces and other technologies for seamlessly transitioning between calls of different types can be implemented. The technologies can be implemented to give the impression of a single call that is upgraded from one call type to another. A new application can register so that an appropriate user interface control appears for activation when seamless call transition is possible. Transitioning for third party applications can thus be supported. Cross-platform implementations can be supported. | 02-19-2015 |
20150382179 | INTELLIGENT COMMUNICATION LINE SELECTION SUPPORTING RICH CONTEXT OF INPUT FACTORS - Various user interfaces and other technologies for interacting with devices that support multiple communication lines can be implemented. For example, intelligent handling of communication line selection can help users deal with multiple communication line scenarios. A user interface for overriding communication line selection can be supported. Users can interact with their devices more efficiently and with less frustration. A wide variety of use default communication line scenarios are supported. Communication lines can be selected based on any of a rich set of input factors. | 12-31-2015 |
Tony Qizhen He, Kirkland, WA US
Patent application number | Description | Published |
---|---|---|
20120323933 | DISPLAYING NOTIFICATIONS BASED ON IMPORTANCE TO THE USER - Embodiments prioritize notifications relating to received communication items for display to a user of a computing device. For each received communication item, the computing device searches for related user data items such as calendar data, contacts, emails, and the like. Each of the user data items has a priority value. The computing device calculates a priority score for the notification based on the priority values associated with the related user data items. The notification is displayed among other notifications based on the calculated priority score relative to the priority scores of the other notifications. In some embodiments, the notifications include an ordered set of notifications displayed as tiles on a user interface of a mobile computing device. | 12-20-2012 |
Wenlei He, Bellevue, WA US
Patent application number | Description | Published |
---|---|---|
20150378697 | INCREMENTAL COMPILATION OF WHOLE PROGRAM CODE - Methods, systems, and computer program products are provided that enable incremental compilation of source code. Attributes of an intermediate language (IL) representation and a compiled representation of a source code are stored. Modified source code that is a revised version of the first source code is received. An IL representation of the modified source code is generated. Attributes of the revised intermediate IL and the stored attributes of the IL representation are compared to determine a first set of functions changed in the modified source code. A second set of functions in the first source code is determined that includes functions affected the determined first set of functions. The first and second sets of functions are compiled to generate a set of compiled functions. Compiled versions of the first and second sets are replaced in the compiled representation of the first source code with the set of compiled functions. | 12-31-2015 |
Xiaodeng He, Issaquah, WA US
Patent application number | Description | Published |
---|---|---|
20140278424 | KERNEL DEEP CONVEX NETWORKS AND END-TO-END LEARNING - Data associated with spoken language may be obtained. An analysis of the obtained data may be initiated for understanding of the spoken language using a deep convex network that is integrated with a kernel trick. The resulting kernel deep convex network may also be constructed by stacking one shallow kernel network over another with concatenation of the output vector of the lower network with the input data vector. A probability associated with a slot that is associated with slot-filling may be determined, based on local, discriminative features that are extracted using the kernel deep convex network. | 09-18-2014 |
Xiaodong He, Issaquah, WA US
Patent application number | Description | Published |
---|---|---|
20080201139 | Generic framework for large-margin MCE training in speech recognition - A method and apparatus for training an acoustic model are disclosed. A training corpus is accessed and converted into an initial acoustic model. Scores are calculated for a correct class and competitive classes, respectively, for each token given the initial acoustic model. Also, a sample-adaptive window bandwidth is calculated for each training token. From the calculated scores and the sample-adaptive window bandwidth values, loss values are calculated based on a loss function. The loss function, which may be derived from a Bayesian risk minimization viewpoint, can include a margin value that moves a decision boundary such that token-to-boundary distances for correct tokens that are near the decision boundary are maximized. The margin can either be a fixed margin or can vary monotonically as a function of algorithm iterations. The acoustic model is updated based on the calculated loss values. This process can be repeated until an empirical convergence is met. | 08-21-2008 |
20090063126 | Validation of the consistency of automatic terminology translation - A method of determining the consistency of training data for a machine translation system is disclosed. The method includes receiving a signal indicative of a source language corpus and a target language corpus. A textual string is extracted from the source language corpus. The textual string is aligned with the target language corpus to identify a translation for the textual string from the target language corpus. A consistency index is calculated based on a relationship between the textual string from the source language corpus and the translation. An indication of the consistency index is stored on a tangible medium. | 03-05-2009 |
20090112573 | Word-dependent transition models in HMM based word alignment for statistical machine translation - A word alignment modeler uses probabilistic learning techniques to train “word-dependent transition models” for use in constructing phrase level Hidden Markov Model (HMM) based word alignment models. As defined herein, “word-dependent transition models” provide a probabilistic model wherein for each source word in training data, a self-transition probability is modeled in combination with a probability of jumping from that particular word to a different word, thereby providing a full transition model for each word in a source phrase. HMM based word alignment models are then used for various word alignment and machine translation tasks. In additional embodiments sparse data problems (i.e., rarely used words) are addressed by using probabilistic learning techniques to estimate word-dependent transition model parameters by maximum a posteriori (MAP) training. | 04-30-2009 |
20090240486 | HMM ALIGNMENT FOR COMBINING TRANSLATION SYSTEMS - A computing system configured to produce an optimized translation hypothesis of text input into the computing system. The computing system includes a plurality of translation machines. Each of the translation machines is configured to produce their own translation hypothesis from the same text. An optimization machine is connected to the plurality of translation machines. The optimization machine is configured to receive the translation hypotheses from the translation machines. The optimization machine is further configured to align, word-to-word, the hypotheses in the plurality of hypotheses by using a hidden Markov model. | 09-24-2009 |
20100161330 | SPEECH MODELS GENERATED USING COMPETITIVE TRAINING, ASYMMETRIC TRAINING, AND DATA BOOSTING - Speech models are trained using one or more of three different training systems. They include competitive training which reduces a distance between a recognized result and a true result, data boosting which divides and weights training data, and asymmetric training which trains different model components differently. | 06-24-2010 |
20100311030 | USING COMBINED ANSWERS IN MACHINE-BASED EDUCATION - Described is a technology for learning a foreign language or other subject. Answers (e.g., translations) to questions (e.g., sentences to translate) received from learners are combined into a combined answer that serves as a representative model answer for those learners. The questions also may be provided to machine subsystems to generate machine answers, e.g., machine translators, with those machine answers used in the combined answer. The combined answer is used to evaluate each learner's individual answer. The evaluation may be used to compute profile information that is then fed back for use in selecting further questions, e.g., more difficult sentences as the learners progress. Also described is integrating the platform/technology into a web service. | 12-09-2010 |
20110307244 | JOINT OPTIMIZATION FOR MACHINE TRANSLATION SYSTEM COMBINATION - A joint optimization strategy is employed for combining translation hypotheses from multiple machine-translation systems. Decisions on word alignment, between the hypotheses, ordering, and selection of a combined translation output are made jointly in accordance with a set of features. Additional features that model alignment and ordering behavior are also provided and utilized. | 12-15-2011 |
20120131031 | DEPENDENCY-BASED QUERY EXPANSION ALTERATION CANDIDATE SCORING - An alteration candidate for a query can be scored. The scoring may include computing one or more query-dependent feature scores and/or one or more intra-candidate dependent feature scores. The computation of the query-dependent feature score(s) can be based on dependencies to multiple query terms from each of one or more alteration terms (i.e., for each of the one or more alteration terms, there can be dependencies to multiple query terms that form at least a portion of the basis for the query-dependent feature score(s)). The computation of the intra-candidate dependent feature score(s) can be based on dependencies between different terms in the alteration candidate. A candidate score can be computed using the query dependent feature score(s) and/or the intra-candidate dependent feature score(s). Additionally, the candidate score can be used in determining whether to select the candidate to expand the query. If selected, the candidate can be used to expand the query. | 05-24-2012 |
20120143591 | INTEGRATIVE AND DISCRIMINATIVE TECHNIQUE FOR SPOKEN UTTERANCE TRANSLATION - Architecture that provides the integration of automatic speech recognition (ASR) and machine translation (MT) components of a full speech translation system. The architecture is an integrative and discriminative approach that employs an end-to-end objective function (the conditional probability of the translated sentence (target) given the source language's acoustic signal, as well as the associated BLEU score in the translation, as a goal in the integrated system. This goal defines the theoretically correct variables to determine the speech translation system output using a Bayesian decision rule. These theoretically correct variables are modified in practical use due to known imperfections of the various models used in building the full speech translation system. The disclosed approach also employs automatic training of these variables using minimum classification error (MCE) criterion. The measurable BLEU scores are used to facilitate the implementation of the MCE training procedure in a step that defines the class-specific discriminant function. | 06-07-2012 |
20120150836 | TRAINING PARSERS TO APPROXIMATELY OPTIMIZE NDCG - A supervised technique uses relevance judgments to train a dependency parser such that it approximately optimizes Normalized Discounted Cumulative Gain (NDCG) in information retrieval. A weighted tree edit distance between the parse tree for a query and the parse tree for a document is added to a ranking function, where the edit distance weights are parameters from the parser. Using parser parameters in the ranking function enables approximate optimization of the parser's parameters for NDCG by adding some constraints to the objective function. | 06-14-2012 |
20120203539 | SELECTION OF DOMAIN-ADAPTED TRANSLATION SUBCORPORA - Architecture that provides the capability to subselect the most relevant data from an out-domain corpus to use either in isolation or in combination conjunction with in-domain data. The architecture is a domain adaptation for machine translation that selects the most relevant sentences from a larger general-domain corpus of parallel translated sentences. The methods for selecting the data include monolingual cross-entropy measure, monolingual cross-entropy difference, bilingual cross entropy, and bilingual cross-entropy difference. A translation model is trained on both the in-domain data and an out-domain subset, and the models can be interpolated together to boost performance on in-domain translation tasks. | 08-09-2012 |
20120254217 | Enhanced Query Rewriting Through Click Log Analysis - Systems, methods, and computer media for identifying related strings for search query rewriting are provided. Session data for a user search query session in an accessed click log data is identified. It is determined whether a first additional search query in the session data is related to a first user search query based on at least one of: dwell time; a number of search result links clicked on; and similarity between web page titles or uniform resource locators (URLs). When related, the first additional search query is incorporated into a list of strings related to the first user search query. One or more supplemental strings that are related to the first user search query are also identified. The identified supplemental strings are also included in the list of strings related to the first user search query. | 10-04-2012 |
20120254218 | Enhanced Query Rewriting Through Statistical Machine Translation - Systems, methods, and computer media for identifying query rewriting replacement terms are provided. A list of related string pairs each comprising a first string and second string is received. The first string of each related string pair is a user search query extracted from user click log data. For one or more of the related string pairs, the string pair is provided as inputs to a statistical machine translation model. The model identifies one or more pairs of corresponding terms, each pair of corresponding terms including a first term from the first string and a second term from the second string. The model also calculates a probability of relatedness for each of the one or more pairs of corresponding terms. Term pairs whose calculated probability of relatedness exceeds a threshold are characterized as query term replacements and incorporated, along with the probability of relatedness, into a query rewriting candidate database. | 10-04-2012 |
20130110491 | DISCRIMINATIVE LEARNING OF FEATURE FUNCTIONS OF GENERATIVE TYPE IN SPEECH TRANSLATION | 05-02-2013 |
20130132083 | GENERIC FRAMEWORK FOR LARGE-MARGIN MCE TRAINING IN SPEECH RECOGNITION - A method and apparatus for training an acoustic model are disclosed. A training corpus is accessed and converted into an initial acoustic model. Scores are calculated for a correct class and competitive classes, respectively, for each token given the initial acoustic model. Also, a sample-adaptive window bandwidth is calculated for each training token. From the calculated scores and the sample-adaptive window bandwidth values, loss values are calculated based on a loss function. The loss function, which may be derived from a Bayesian risk minimization viewpoint, can include a margin value that moves a decision boundary such that token-to-boundary distances for correct tokens that are near the decision boundary are maximized. The margin can either be a fixed margin or can vary monotonically as a function of algorithm iterations. The acoustic model is updated based on the calculated loss values. This process can be repeated until an empirical convergence is met. | 05-23-2013 |
20130311504 | DEPENDENCY-BASED QUERY EXPANSION ALTERATION CANDIDATE SCORING - An alteration candidate for a query can be scored. The scoring may include computing one or more query-dependent feature scores and/or one or more intra-candidate dependent feature scores. The computation of the query-dependent feature score(s) can be based on dependencies to multiple query terms from each of one or more alteration terms (i.e., for each of the one or more alteration terms, there can be dependencies to multiple query terms that form at least a portion of the basis for the query-dependent feature score(s)). The computation of the intra-candidate dependent feature score(s) can be based on dependencies between different terms in the alteration candidate. A candidate score can be computed using the query dependent feature score(s) and/or the intra-candidate dependent feature score(s). Additionally, the candidate score can be used in determining whether to select the candidate to expand the query. If selected, the candidate can be used to expand the query. | 11-21-2013 |
20140279995 | QUERY SIMPLIFICATION - Methods, systems, and computer-readable media for query simplification are provided. A search engine executed by a server receives a query. In response, the search engine determines whether the query is a long or hard query. For long or hard queries, the search engine drops one or more terms based on search engine logs. The search engine may utilize statistical models like machine translation, condition random fields, or max entropy, to identify the terms that should be dropped. The search engine obtains search results for the simplified query and transmits the results to a user that provided the query. | 09-18-2014 |
20140365201 | TRAINING MARKOV RANDOM FIELD-BASED TRANSLATION MODELS USING GRADIENT ASCENT - Various technologies described herein pertain to training and utilizing a general, statistical framework for modeling translation via Markov random fields (MRFs). An MRF-based translation model can be employed in a statistical machine translation (SMT) system. The MRF-based translation model allows for arbitrary features extracted from a phrase pair to be incorporated as evidence. The parameters of the model are estimated using a large-scale discriminative training approach based on stochastic gradient ascent and an N-best list based expected Bilingual Evaluation Understudy (BLEU) as an objective function. | 12-11-2014 |
20150066496 | ASSIGNMENT OF SEMANTIC LABELS TO A SEQUENCE OF WORDS USING NEURAL NETWORK ARCHITECTURES - Technologies pertaining to slot filling are described herein. A deep neural network, a recurrent neural network, and/or a spatio-temporally deep neural network are configured to assign labels to words in a word sequence set forth in natural language. At least one label is a semantic label that is assigned to at least one word in the word sequence. | 03-05-2015 |
20150074027 | Deep Structured Semantic Model Produced Using Click-Through Data - A deep structured semantic module (DSSM) is described herein which uses a model that is discriminatively trained based on click-through data, e.g., such that a conditional likelihood of clicked documents, given respective queries, is maximized, and a condition likelihood of non-clicked documents, given the queries, is reduced. In operation, after training is complete, the DSSM maps an input item into an output item expressed in a semantic space, using the trained model. To facilitate training and runtime operation, a dimensionality-reduction module (DRM) can reduce the dimensionality of the input item that is fed to the DSSM. A search engine may use the above-summarized functionality to convert a query and a plurality of documents into the common semantic space, and then determine the similarity between the query and documents in the semantic space. The search engine may then rank the documents based, at least in part, on the similarity measures. | 03-12-2015 |
20150278200 | Convolutional Latent Semantic Models and their Applications - Functionality is described herein for transforming first and second symbolic linguistic items into respective first and second continuous-valued concept vectors, using a deep learning model, such as a convolutional latent semantic model. The model is designed to capture both the local and global linguistic contexts of the linguistic items. The functionality then compares the first concept vector with the second concept vector to produce a similarity measure. More specifically, the similarity measure expresses the closeness between the first and second linguistic items in a high-level semantic space. In one case, the first linguistic item corresponds to a query, and the second linguistic item may correspond to a phrase, or a document, or a keyword, or an ad, etc. In one implementation, the convolutional latent semantic model is produced in a training phase based on click-through data. | 10-01-2015 |
20150293976 | Context-Sensitive Search Using a Deep Learning Model - A search engine is described herein for providing search results based on a context in which a query has been submitted, as expressed by context information. The search engine operates by ranking a plurality of documents based on a consideration of the query, and based, in part, on a context concept vector and a plurality of document concept vectors, both generated using a deep learning model (such as a deep neural network). The context concept vector is formed by a projection of the context information into a semantic space using the deep learning model. Each document concept vector is formed by a projection of document information, associated with a particular document, into the same semantic space using the deep learning model. The ranking operates by favoring documents that are relevant to the context within the semantic space, and disfavoring documents that are not relevant to the context. | 10-15-2015 |
20150363688 | MODELING INTERESTINGNESS WITH DEEP NEURAL NETWORKS - An “Interestingness Modeler” uses deep neural networks to learn deep semantic models (DSM) of “interestingness.” The DSM, consisting of two branches of deep neural networks or their convolutional versions, identifies and predicts target documents that would interest users reading source documents. The learned model observes, identifies, and detects naturally occurring signals of interestingness in click transitions between source and target documents derived from web browser logs. Interestingness is modeled with deep neural networks that map source-target document pairs to feature vectors in a latent space, trained on document transitions in view of a “context” and optional “focus” of source and target documents. Network parameters are learned to minimize distances between source documents and their corresponding “interesting” targets in that space. The resulting interestingness model has applicable uses, including, but not limited to, contextual entity searches, automatic text highlighting, prefetching documents of likely interest, automated content recommendation, automated advertisement placement, etc. | 12-17-2015 |
Yaming He, Redmond, WA US
Patent application number | Description | Published |
---|---|---|
20100316126 | MOTION BASED DYNAMIC RESOLUTION MULTIPLE BIT RATE VIDEO ENCODING - A video encoding system encodes video streams for multiple bit rate video streaming using an approach that permits the encoded resolution to vary based, at least in part, on motion complexity. The video encoding system dynamically decides an encoding resolution for segments of the multiple bit rate video streams that varies with video complexity so as to achieve a better visual experience for multiple bit rate streaming. Motion complexity may be considered separately, or along with spatial complexity, in making the resolution decision. | 12-16-2010 |
20110305273 | PARALLEL MULTIPLE BITRATE VIDEO ENCODING - A multiple bitrate (MBR) video encoding management tool utilizes available processing units for parallel MBR video encoding. For example, instead of focusing only on multi-threading of encoding tasks for a single picture or group of pictures (GOP), the management tool parallelizes the encoding of multiple GOPs between different processing units and/or different computing systems. With this parallel MBR video encoding architecture, different GOPs can be encoded in parallel. To facilitate such parallel encoding, data dependencies between GOPs are removed. The management tool can adjust the number of GOPs to encode in parallel on a computing system so as to favor parallelism of encoding for different GOPs at the expense of parallelism of encoding inside a GOP, or vice versa, and thereby set a suitable balance between encoding latency and throughput. | 12-15-2011 |
20120106806 | Face Recognition in Video Content - The subject disclosure relates to face recognition in video. Face detection data in frames of input data are used to generate face galleries, which are labeled and used in recognizing faces throughout the video. Metadata that associates the video frame and the face are generated and maintained for subsequent identification. Faces other than those found by face detection may be found by face tracking, in which facial landmarks found by the face detection are used to track a face over previous and/or subsequent video frames. Once generated, the maintained metadata may be accessed to efficiently determine the identity of a person corresponding to a viewer-selected face. | 05-03-2012 |
20120281886 | STAGED ELEMENT CLASSIFICATION - Various examples are disclosed herein that relate to staged element classification. For example, one disclosed example provides a method of classifying elements by forming elements for classification into a plurality of first-level sets in a first stage, generating primary groups within the first-level sets based on element similarity, forming a plurality of second-level sets from the first-level sets in a second stage, generating secondary groups within the second-level sets based on element similarity, and merging a plurality of the primary and/or secondary groups based on element similarity. | 11-08-2012 |
20130215977 | SCALABLE MULTI-THREAD VIDEO DECODING - Decoding tasks are identified for decoding encoded video. Decoding tasks may include entropy decoding tasks, motion compensation tasks, inverse frequency transform tasks, inverse quantization tasks, intra decoding tasks, loop filtering tasks, or other tasks. Task dependencies are identified for the video decoding tasks. For example, one or more decoding tasks may depend on prior completion of entropy decoding tasks. The decoding tasks are prioritized based at least in part on the task dependencies. For example, a higher priority may be assigned to tasks that must be completed before other tasks that depend on them can begin. Prioritized decoding tasks are selected to be performed by hardware threads. For example, a first hardware thread may perform a first decoding task that does not depend on any uncompleted tasks while a second hardware thread performs a second decoding task that does not depend on any uncompleted tasks. | 08-22-2013 |
20140233652 | SCALABLE MULTI-THREAD VIDEO DECODING - Decoding tasks are identified for decoding encoded video. Decoding tasks may include entropy decoding tasks, motion compensation tasks, inverse frequency transform tasks, inverse quantization tasks, intra decoding tasks, loop filtering tasks, or other tasks. Task dependencies are identified for the video decoding tasks. For example, one or more decoding tasks may depend on prior completion of entropy decoding tasks. The decoding tasks are prioritized based at least in part on the task dependencies. For example, a higher priority may be assigned to tasks that must be completed before other tasks that depend on them can begin. Prioritized decoding tasks are selected to be performed by hardware threads. For example, a first hardware thread may perform a first decoding task that does not depend on any uncompleted tasks while a second hardware thread performs a second decoding task that does not depend on any uncompleted tasks. | 08-21-2014 |
Yeye He, Redmond, WA US
Patent application number | Description | Published |
---|---|---|
20150154681 | ENRICHING PRODUCT CATALOG WITH PRODUCT NAME KEYWORDS - A keyword generator identifies words or phrases of interest in a product catalog and also identifies synonyms for the words or phrases of interest. The synonyms are integrated into the product catalog to generate an enriched product catalog. The level of co-occurrence of synonyms between sets of product catalog entries is identified and, if it meets a threshold level, the product names from the catalog entries are integrated into the other catalog entries in the set, as synonyms. | 06-04-2015 |
20150154682 | ENRICHING PRODUCT CATALOG WITH SEARCH KEYWORDS - A keyword generator identifies words or phrases of interest in a product catalog and also identifies synonyms for the words or phrases of interest. The synonyms are integrated into the product catalog to generate an enriched product catalog. The enriched product catalog is published for use in one or more commercial channels. | 06-04-2015 |
Yin He, Redmond, WA US
Patent application number | Description | Published |
---|---|---|
20140172813 | PROCESSING USER LOG SESSIONS IN A DISTRIBUTED SYSTEM - Systems, methods, and computer media for efficiently processing user log data are provided. The log data is progressively processed in variable sized windows based on a specified time period. The log data may be anonymized to protect user privacy. A log server processes the windowed log data in phases. The first phase includes fast data like page view log data. Subsequent phases include slow data like session data which may build on the page view data processed in the first phase. The log server identifies metrics based on the log data processed at each phase. Based on the identified metrics, the log server may identify interests across a community of users or for specific users. | 06-19-2014 |
Yiwu He, Mercer Island, WA US
Patent application number | Description | Published |
---|---|---|
20110245092 | DIAGNOSING AND MONITORING DEPRESSION DISORDERS BASED ON MULTIPLE SERUM BIOMARKER PANELS - Materials and Methods related to developing a unipolar depression (MDD) disease score in a subject using a multi-parameter system to measure a plurality of parameters, and an algorithm to calculate a score. | 10-06-2011 |
20110269633 | INFLAMMATORY BIOMARKERS FOR MONITORING DEPRESSIVE DISORDERS - Materials and Methods related to developing a unipolar depression (MDD) disease score in a subject using a multi-parameter system to measure a plurality of parameters, and an algorithm to calculate a score. | 11-03-2011 |
20120289422 | QUANTITATIVE DIAGNOSTIC METHODS USING MULTIPLE PARAMETERS - Materials and Methods related to diagnosing a clinical condition in a subject, or determining the subject's predisposition to develop the clinical condition, using a multi-parameter system to measure a plurality of parameters and an algorithm to determine a disease score. | 11-15-2012 |
20140135224 | QUANTITATIVE DIAGNOSTIC METHODS USING MULTIPLE PARAMETERS - Materials and methods related to diagnosing a clinical condition in a subject, or determining the subject's predisposition to develop the clinical condition, using a multi-parameter system to measure a plurality of parameters and an algorithm to determine a disease score. | 05-15-2014 |
20140257708 | DIAGNOSING AND MONITORING DEPRESSION DISORDERS - Materials and methods related to developing a disease score for a depression disorder (e.g., unipolar depression or major depressive disorder) in a subject using a multi-parameter system to measure a plurality of parameters, and an algorithm to calculate the score. The materials and methods can be used to, for example, diagnose depression disorders, or determine a subject's predisposition to develop a depression disorder. The methods also can include using a multi-parameter hypermapping system and algorithms related thereto. | 09-11-2014 |
Yudong He, Kirkland, WA US
Patent application number | Description | Published |
---|---|---|
20090157326 | Diagnosis and prognosis of breast cancer patients - The present invention relates to genetic markers whose expression is correlated with breast cancer. Specifically, the invention provides sets of markers whose expression patterns can be used to differentiate clinical conditions associated with breast cancer, such as the presence or absence of the estrogen receptor ESR1, and BRCA1 and sporadic tumors, and to provide information on the likelihood of tumor distant metastases within five years of initial diagnosis. The invention relates to methods of using these markers to distinguish these conditions. The invention also relates to kits containing ready-to-use microarrays and computer software for data analysis using the statistical methods disclosed herein. | 06-18-2009 |
20090204333 | METHODS FOR USING CO-REGULATED GENESETS TO ENHANCE DETECTION AND CLASSIFICATION OF GENE EXPRESSION PATTERNS - The present invention provides methods for enhanced detection of biological response patterns. In one embodiment of the invention, genes are grouped into basis genesets according to the co-regulation of their expression. Expression of individual genes within a geneset is indicated with a single gene expression value for the geneset by a projection process. The expression values of genesets, rather than the expression of individual genes, are then used as the basis for comparison and detection of biological response with greatly enhanced sensitivity. In another embodiment of the invention, biological responses are grouped according to the similarity of their biological profile. | 08-13-2009 |
20090239214 | Prognosis of breast cancer patients - The present invention relates to sets of genetic markers whose expression is correlated with prognosis of breast cancer in individuals having breast cancer. Specifically, the invention provides sets of markers whose expression patterns can be used to differentiate individuals having a good prognosis, e.g., no reoccurrence or metastasis within five years of initial diagnosis, and individuals having a poor prognosis, e.g., reoccurrence or metastasis within five years of initial diagnosis. The invention relates to methods of prognosis using these markers. The invention also relates to microarrays containing probes to these markers, and to kits containing ready-to-use microarrays and computer software for data analysis using the prognostic and statistical methods disclosed herein. | 09-24-2009 |
20110301048 | DIAGNOSIS AND PROGNOSIS OF BREAST CANCER PATIENTS - The present invention relates to genetic markers whose expression is correlated with breast cancer. Specifically, the invention provides sets of markers whose expression patterns can be used to differentiate clinical conditions associated with breast cancer, such as the presence or absence of the estrogen receptor ESR1, and BRCA1 and sporadic tumors, and to provide information on the likelihood of tumor distant metastases within five years of initial diagnosis. The invention relates to methods of using these markers to distinguish these conditions. The invention also relates to kits containing ready-to-use microarrays and computer software for data analysis using the statistical methods disclosed herein. | 12-08-2011 |
20130116145 | DIAGNOSIS AND PROGNOSIS OF BREAST CANCER PATIENTS - The present invention relates to genetic markers whose expression is correlated with breast cancer. Specifically, the invention provides sets of markers whose expression patterns can be used to differentiate clinical conditions associated with breast cancer, such as the presence or absence of the estrogen receptor ESR1, and BRCA1 and sporadic tumors, and to provide information on the likelihood of tumor distant metastases within five years of initial diagnosis. The invention relates to methods of using these markers to distinguish these conditions. The invention also relates to kits containing ready-to-use microarrays and computer software for data analysis using the statistical methods disclosed herein. | 05-09-2013 |
Yuxiong He, Bellevue, WA US
Patent application number | Description | Published |
---|---|---|
20120324466 | Scheduling Execution Requests to Allow Partial Results - The subject disclosure is directed towards scheduling requests using quality values that are defined for partial responses to the requests. For each request in a queue, an associated processing time is determined using a system load and/or the quality values. The associated processing time is less than or equal to a service demand, which represents an amount of time to produce a complete response. | 12-20-2012 |
20140172914 | GRAPH QUERY PROCESSING USING PLURALITY OF ENGINES - Graph queries are processed using a plurality of independent query execution engines. A graph query submitted to a graph database which is modeled by an attributed graph is received. The graph query is decomposed into a plurality of query components. For each of the query components, a one of the query execution engines that is available to process the query component is identified, a sub-query representing the query component is generated, the sub-query is sent to the identified query execution engine for processing, and results for the sub-query are received from the identified query execution engine. The results received are then combined to generate a response to the graph query. | 06-19-2014 |
20140181825 | Assigning Jobs to Heterogeneous Processing Modules - A processing system is described which assigns jobs to heterogeneous processing modules. The processing system assigns jobs to the processing modules in a manner that attempts to accommodate the service demands of the jobs, but without advance knowledge of the service demands. In one case, the processing system implements the processing modules as computing units that have different physical characteristics. Alternatively, or in addition, the processing system may implement the processing modules as threads that are executed by computing units. Each thread which runs on a computing unit offers a level of performance that depends on a number of other threads that are simultaneously being executed by the same computing unit. | 06-26-2014 |
Zifeng He, Redmond, WA US
Patent application number | Description | Published |
---|---|---|
20080222598 | Using a system of annotations to generate views and adapters - Various technologies and techniques are disclosed for using annotations in contracts to generate views and adapters. A framework is provided that supports annotations that can be used in a contract to allow a user to express how one or more members of the contract should be represented in at least one view, and to express how to adapt the members in at least one adapter. Annotations can also be used to indicate which of the components should be programmatically generated. The contract is accessed and at least a portion of the at least one view and the at least one adapter are generated based on the annotations and other information contained in the contract. For example, source code can programmatically be generated for a host side view, an add-in side view, a host side adapter, and an add-in side adapter using the various contract details and the annotations. | 09-11-2008 |
20090006629 | Instantiating a communication pipeline between software - An embodiment not only instantiates a pipeline that allows communication between first and second executable nodes regardless of whether the nodes have compatible or incompatible interface views, but also structures the pipeline based on the relationship between the interface views so, for example, the pipeline does not consume unnecessary overhead or introduce unnecessary data-transfer delay. For example, an embodiment compares a first interface view of a first executable node to a second interface view of a second executable node. A first communication pipeline having a first structure is instantiated between the first and second nodes if the first view has a first pre-established relationship to the second view, and a second communication pipeline having a second structure is instantiated between the first and second nodes if the first view has a second pre-established relationship to the second view. | 01-01-2009 |