Abstract: Statistical Machine Translation (SMT) based search query spelling correction techniques are described herein. In one or more implementations, search data regarding searches performed by clients may be logged. The logged data includes query correction pairs that may be used to ascertain error patterns indicating how misspelled substrings may be translated to corrected substrings. The error patterns may be used to determine suggestions for an input query and to develop query correction models used to translate the input query to a corrected query. In one or more implementations, probabilistic features from multiple query correction models are combined to score different correction candidates. One or more top scoring correction candidates may then be exposed as suggestions for selection by a user and/or provided to a search engine to conduct a corresponding search using the corrected query version(s).

Abstract: A method to execute computer-actionable directives conveyed in human speech comprises: receiving audio data recording speech from one or more speakers; converting the audio data into a linguistic representation of the recorded speech; detecting a target corresponding to the linguistic representation; committing to the data structure language data associated with the detected target and based on the linguistic representation; parsing the data structure to identify one or more of the computer-actionable directives; and submitting the one or more of the computer-actionable directives to the computer for processing.

Abstract: A method for use with a computing device is provided. The method may include executing one or more programs of an intelligent digital assistant system at a processor and presenting a user interface to a user. At the processor, the method may include receiving natural language user input from the user, parsing the user input at an intent handler to determine an intent template with slots, populating the slots in the intent template with information from user input, and performing resolution on the intent template to partially resolve unresolved information. If a slot with missing slot information exists in the partially resolved intent template, a loop may be executed at the processor to fill the slots. The method may include, at the processor, determining that all required information is available and resolved and generating a rule based upon the intent template with all required information being available and resolved.

Abstract: Intelligent assistant systems, methods and computing devices are disclosed for training a machine learning-based parser to derive user intents. A method comprises analyzing with a feeder parser a surface form of a user input. A user intent underlying the surface form is derived by the feeder parser. The surface form and the user intent are provided to a machine learning-based parser and used to enhance a training set of the machine learning-based parser.

Abstract: Long short term memory units that accept a non-predefined number of inputs are used to provide natural language relation extraction over a user-specified range on content. Content written for human consumption is parsed with distant supervision in segments (e.g., sentences, paragraphs, chapters) to determine relationships between various words within and between those segments.

Abstract: Examples are generally directed towards automatic assessment of machine generated conversational responses. Context-message-response n-tuples are extracted from at least one source of conversational data to generate a set of multi-reference responses. A response in the set of multi-reference responses includes it context-message data pair and rating. The rating indicates a quality of the response relative to the context-message data pair. A response assessment engine generates a metric score for a machine-generated response based on an assessment metric and the set of multi-reference responses. The metric score indicates a quality of the machine-generated conversational response relative to a user-generated message and a context of the user-generated message. A response generation system of a computing device, such as a digital assistant, is optimized and adjusted based on the metric score to improve the accuracy, quality, and relevance of responses output to the user.

Abstract: Systems, methods, and computer-readable media for providing entity relation extraction across sentences in a document using distant supervision. In some examples, a computing device can receive an input, such as a document comprising a plurality of sentences. The computing device can identify syntactic and/or semantic links between words in a sentence and/or between words in different sentences, and extract relationships between entities throughout the document. Techniques and technologies described herein populate a knowledge base (e.g., a table, chart, database etc.) of entity relations based on the extracted relationships. An output of the populated knowledge base can be used by a classifier to identify additional relationships between entities in various documents. Example techniques described herein can apply machine learning to train the classifier to predict relations between entities. The classifier can be trained using known entity relations, syntactic links and/or semantic links.

Abstract: Examples are generally directed towards automatic assessment of machine generated conversational responses. Context-message-response n-tuples are extracted from at least one source of conversational data to generate a set of multi-reference responses. A response in the set of multi-reference responses includes it context-message data pair and rating. The rating indicates a quality of the response relative to the context-message data pair. A response assessment engine generates a metric score for a machine-generated response based on an assessment metric and the set of multi-reference responses. The metric score indicates a quality of the machine-generated conversational response relative to a user-generated message and a context of the user-generated message. A response generation system of a computing device, such as a digital assistant, is optimized and adjusted based on the metric score to improve the accuracy, quality, and relevance of responses output to the user.

Abstract: Examples described herein dynamically personalize a digital assistant for a specific user, creating a personal connection between the digital assistant and the user. The digital assistant accesses user activity and generates queries based on the user activity. The digital assistant facilitates natural language conversations as machine learning sessions between the digital assistant and the user using the one or more queries to learn the user's preferences and receives user input from the user during the learning session in response to the queries. The digital assistant dynamically updates a personalized profile for the user based on the user input during the natural language conversations.

Abstract: The claimed subject matter provides systems and/or methods for training feature weights in a statistical machine translation model. The system can include components that obtain lists of translation hypotheses and associated feature values, set a current point in the multidimensional feature weight space to an initial value, chooses a line in the feature weight space that passes through the current point, and resets the current point to optimize the feature weights with respect to the line. The system can further include components that set the current point to be a best point attained, reduce the list of translation hypotheses based on a determination that a particular hypothesis has never been touched in optimizing the feature weights from at least one of an initial staring point or a randomly selected restarting point, and output the point ascertained to be the best point in the feature weight space.

Abstract: The universal text input technique described herein addresses the difficulties of typing text in various languages and scripts, and offers a unified solution, which combines character conversion, next word prediction, spelling correction and automatic script switching to make it extremely simple to type any language from any device. The technique provides a rich and seamless input experience in any language through a universal IME (input method editor). It allows a user to type in any script for any language using a regular qwerty keyboard via phonetic input and at the same time allows for auto-completion and spelling correction of words and phrases while typing. The technique also provides a modeless input that automatically turns on and off an input mode that changes between different types of script.

Abstract: Systems and methods for automatically extracting parallel word sequences from comparable corpora are described. Electronic documents, such as web pages belonging to a collaborative online encyclopedia, are analyzed to locate parallel word sequences between electronic documents written in different languages. These parallel word sequences are then used to train a machine translation system that can translate text from one language to another.

Abstract: Statistical Machine Translation (SMT) based search query spelling correction techniques are described herein. In one or more implementations, search data regarding searches performed by clients may be logged. The logged data includes query correction pairs that may be used to ascertain error patterns indicating how misspelled substrings may be translated to corrected substrings. The error patterns may be used to determine suggestions for an input query and to develop query correction models used to translate the input query to a corrected query. In one or more implementations, probabilistic features from multiple query correction models are combined to score different correction candidates. One or more top scoring correction candidates may then be exposed as suggestions for selection by a user and/or provided to a search engine to conduct a corresponding search using the corrected query version(s).

Abstract: Various technologies described herein pertain to detecting machine translated content. Documents in a document pair are mutual lingual translations of each other. Further, document level features of the documents in the document pair can be identified. The document level features can correlate with translation quality between the documents in the document pair. Moreover, statistical classification can be used to detect whether the document pair is generated through machine translation based at least in part upon the document level features. Further, a first document can be a machine translation of a second document in the document pair or a disparate document when generated through machine translation.

Abstract: The universal text input technique described herein addresses the difficulties of typing text in various languages and scripts, and offers a unified solution, which combines character conversion, next word prediction, spelling correction and automatic script switching to make it extremely simple to type any language from any device. The technique provides a rich and seamless input experience in any language through a universal IME (input method editor). It allows a user to type in any script for any language using a regular qwerty keyboard via phonetic input and at the same time allows for auto-completion and spelling correction of words and phrases while typing. The technique also provides a modeless input that automatically turns on and off an input mode that changes between different types of script.

Abstract: Human computation games are provided wherein a player is shown a page, such as a web page. The player is then asked to provide one or more terms that are intended to cause a search engine to return the page in response to performing a query using the terms. The terms provided by the player during game play are then collected, stored, and utilized to improve the performance of the search engine.

Abstract: Systems and methods for automatically extracting parallel word sequences from comparable corpora are described. Electronic documents, such as web pages belonging to a collaborative online encyclopedia, are analyzed to locate parallel word sequences between electronic documents written in different languages. These parallel word sequences are then used to train a machine translation system that can translate text from one language to another.

Abstract: Human computation games are provided wherein a player is shown a page, such as a web page. The player is then asked to provide one or more terms that are intended to cause a search engine to return the page in response to performing a query using the terms. The terms provided by the player during game play are then collected, stored, and utilized to improve the performance of the search engine.

Abstract: The claimed subject matter provides systems and/or methods for training feature weights in a statistical machine translation model. The system can include components that obtain lists of translation hypotheses and associated feature values, set a current point in the multidimensional feature weight space to an initial value, chooses a line in the feature weight space that passes through the current point, and resets the current point to optimize the feature weights with respect to the line. The system can further include components that set the current point to be a best point attained, reduce the list of translation hypotheses based on a determination that a particular hypothesis has never been touched in optimizing the feature weights from at least one of an initial staring point or a randomly selected restarting point, and output the point ascertained to be the best point in the feature weight space.

Abstract: The claimed subject matter provides systems and/or methods that minimize error rate training for statistical machine translation. The systems can include devices that optimize a statistical machine translation model for translating between a first natural language and a second natural language by generating lists of n-best translation hypotheses and associated feature weights, optimizing the associated feature weights with respect to the lists of n-best translation hypotheses, and thereafter determining a translation quality measurement for the training sets from which the lists of n-best translation hypotheses were derived.