Amazon Technologies, Inc.

Silverstream Software, Inc.

Oracle International Corporation

Novell, Inc.

Endeca Technologies, Inc.

Alexander David Rosen

Paul H DiCristina

Hugh Evan Secker-Walker

Kenneth John Basye

Michael James Rodehorst

Ariya Rastrow

Baiyang Liu

Vladimir V Zelevinsky

Adam J Ferrari

Jason Furtado

Naresh Narayanan

Andrew Volpe

Gautam Tiwari

Omri Traub

Spiro Michaylov

Aaron Lee Mathers Challenner

Joshua William Kapell

Daniel Robert Rashid

John Huffaker

Shaun Nidhiri Joseph

Matthew L Brandwein

Colin Wills Wightman

Jake Kramer

Benjamin S Scarlet

George Jay Tucker

Michael Maximilian Emanuel Bisani

Sonal Pareek

Yuwang Yin

Eli Daniel

Gregory C Scott

Eugene Jitomirsky

Vladimir Gluzman Peregrine

Systems and methods for generating sound profiles of artificial commands detected by multiple voice activated electronic devices is described herein. In some embodiments, numerous voice activated electronic devices may send audio data representing a phrase to a backend system at a substantially same time. Text data representing the phrase, and counts for instances of that text data, may be generated. If the number of counts exceeds a predefined threshold, the backend system may cause any remaining response generation functionality that particular command that is in excess of the predefined threshold to be stopped, and those devices returned to a sleep state. In some embodiments, a sound profile unique to the phrase that caused the excess of the predefined threshold may be generated such that future instances of the same phrase may be recognized prior to text data being generated, conserving the backend system's resources.

Sound profile generation based on speech recognition results exceeding a threshold

User-specific language models (LMs) that include internal word indexes to a word table specific to the user-specific LM rather than a word table specific to a system-wide LM. When the system-wide LM is updated, the word table of the user-specific LM may be updated to translate the user-specific indices to system-wide indices. This prevents having to update the internal indices of the user-specific LM every time the system-wide LM is updated.

Customized speech processing language models

Systems and methods for detecting similar audio being received by separate voice activated electronic devices, and ignoring those commands, is described herein. In some embodiments, a voice activated electronic device may be activated by a wakeword that is output by the additional electronic device, such as a television or radio, may capture audio of sound subsequently following the wakeword, and may send audio data representing the sound to a backend system. Upon receipt, the backend system may, in parallel to performing automated speech recognition processing to the audio data, generate a sound profile of the audio data, and may compare that sound profile to sound profiles of recently received audio data and/or flagged sound profiles. If the generated sound profile is determined to match another sound profiles, then the automated speech recognition processing may be stopped, and the voice activated electronic device may be instructed to return to a keyword spotting mode. If the matching sound profile is not already stored in a database of known sound profiles, it can be stored for future comparisons.

Methods and devices for ignoring similar audio being received by a system

In a dynamic automatic speech recognition (ASR) processing system, ASR processing may be configured to estimate a latency of returning speech results to a user based on work being done by an ASR processor. The ASR processing system may measure work done by an ASR processor by measuring one or more time independent metrics and comparing the metrics to threshold values. If the metrics exceed the thresholds, the ASR system may take steps to reduce latency associated with processing the utterance, including adjusting a speech recognition parameter.

Dynamic pruning in speech recognition

In an automatic speech recognition (ASR) processing system, ASR processing may be configured to reduce a latency of returning speech results to a user. The latency may be determined by comparing a time stamp of an utterance in process to a current time. Latency may also be estimated based on an endpoint of the utterance or other considerations such as how difficult the utterance may be to process. To improve latency the ASR system may be configured to adjust various processing parameters, such as graph pruning factors, path weights, ASR models, etc. Latency checks and corrections may occur dynamically for a particular utterance while it is being processed, thus allowing the ASR system to adjust to rapidly changing latency conditions.

Reducing speech recognition latency

Determining the end of an utterance for purposes of automatic speech recognition (ASR) may be improved with a system that provides early results and/or incorporates semantic tagging. Early ASR results of an incoming utterance may be prepared based at least in part on an estimated endpoint and processed by a natural language understanding (NLU) process while final results, based at least in part on a final endpoint, are determined. If the early results match the final results, the early NLU results are already prepared for early execution. The endpoint may also be determined based at least in part on the content of the utterance, as represented by semantic tagging output from ASR processing. If the tagging indicate completion of a logical statement, an endpoint may be declared, or a threshold for silent frames prior to declaring an endpoint may be adjusted.

Enhanced endpoint detection for speech recognition

Features are disclosed for identifying and providing command suggestions during automatic speech recognition. As utterances are interpreted, suggestions may be provided based on even partial interpretations to guide users of a client device to commands available via speech recognition.

Command suggestions during automatic speech recognition

Systems and methods for information retrieval are provided that permit users and/or processing entities to access and define synthetic data, synthetic objects, and/or synthetic groupings of data in one or more collections of information. In one embodiment, data access on an information retrieval system can occur through an interpretation layer which interprets any synthetic data against data physically stored in the collection. Synthetic data can define virtual data objects, virtual data elements, virtual data attributes, virtual data groupings, and/or data entities that can be interpreted against data that may be stored physically in the collection of information. The system and methods for information retrieval can return results from the one or more collections of information based not only on the data stored, but also on the virtual data generated from interpretation of the stored data.

Information retrieval and navigation using a semantic layer

A method is disclosed for modifying the results of a search performed in a collection of items by a search and navigation system. The method includes receiving a query from a user interface and determining a navigation state, defined by expressions of attribute-value pairs, based on the received query. The user interface accepts both selecting and deselecting of any of the attribute-value pairs in an expression corresponding to a navigation state to obtain an expression corresponding to a different navigation state, and each selection and deselection forms a new query. The method further includes retrieving, from the collection, items associated with the navigation state to form a set of unmodified search results, the set of unmodified search results having an arrangement for presentation to the user. A rule filter that includes a metadata expression is applied to a set of rules, each rule having a trigger, an action, and metadata. The application of the rule filter to the set of rules includes evaluating the metadata expression of the rule filter based on the metadata of each rule and passing rules for which the metadata expression of the rule filter evaluates as logical true. The trigger of each rule passed by the rule filter is evaluated, and the action of each rule for which the trigger of the rule evaluates as logical true is executed to modify the unmodified search results to form modified search results.

System and method for filtering rules for manipulating search results in a hierarchical search and navigation system

Embodiments of the invention comprise a design tool for designing an application interface. The design tool includes a graphical user interface (GUI) that visually represents a hierarchy of data and the relationships between the data. Thus, the design tool eliminates the need for an interface designer to have independent knowledge of the structure of the data (i.e., the data fields and relationships between the data). The design tool's GUI represents the data and the relationships between the data in a hierarchical display referred to as a data palette. An output hierarchy comprised of output levels is created as the user selects fields from the data palette to be displayed in the application's interface. When a data field is selected, the design tool automatically determines the appropriate interface component and output level of the output hierarchy using the relationships defined for the data. Output levels are associated with interface components that comprise the application's interface. Examples of interface components that can be used in the application's interface include a form, view and subband.

Method and apparatus for building an application interface

Growing community of inventors

Somerville, MA, United States of America

Alexander David Rosen