Microsoft Technology Licensing, LLC

Eric I-Chao Chang

Jian-Lai Zhou

Gary William Flake

Yan Xu

Zhuowen Tu

Junyan Zhu

Abhiram G Khune

Oscar E Murillo

Brett David Brewer

Jonathan C Cluts

Janet Ellen Galore

Karim Taji Farouki

Ali M Vassigh

Jonathan Alan Handler

Raman Chandrasekar

Ian Sands

Flora P Goldthwaite

Sven Pleyer

Yichen Wei

Michael H Tsang

Frank Seide

Darrell Leroy Blegen

Michael Todd Gillam

Tian Bai

Craig Frederick Feied

Peter Burchhardt

Tim Tian Chen

Jiajun Wu

Peter B Thompson

The techniques and systems described herein create and train a multiple clustered instance learning (MCIL) model based on image features and patterns extracted from training images. The techniques and systems separate each of the training images into a plurality of instances (or patches), and then learn multiple instance-level classifiers based on the extracted image features. The instance-level classifiers are then integrated into the MCIL model so that the MCIL model, when applied to unclassified images, can perform image-level classification, patch-level clustering, and pixel-level segmentation.

Multiple cluster instance learning for image classification

Techniques for unsupervised object class discovery via bottom-up multiple class learning are described. These techniques may include receiving multiple images containing one or more object classes. The multiple images may be analyzed to extract top saliency instances and least saliency instances. These saliency instances may be clustered to generate and/or update statistical models. The statistical models may be used to discover the one or more object classes. In some instances, the statistical models may be used to discover object classes of novel images.

Unsupervised object class discovery via bottom up multiple class learning

The claimed subject matter relates to an architecture that can enhance an experience associated with indicia related to a local environment. In particular, the architecture can receive an image that depicts a view of the local environment including a set of entities represented in the image. One or more of the entities can be matched or correlated to modeled entities included in a geospatial model of the environment, potentially based upon location and direction, in order to scope or frame the view depicted in the image to a modeled view. In addition, the architecture can select additional content that can be presented. The additional content typically relates to services or data associated with modeled entities included in the geospatial model or associated with modeled entities included in an image-based data store.

Augmenting a field of view

The claimed subject matter relates to an architecture that can leverage private affinities in order to facilitate or enrich relationships between people. In particular, the architecture can receive a profile associated with a user wherein the profile includes a set of private affinities that are cryptographically protected from public inspection. The architecture can decrypt and/or cryptographically compare a private affinity from the profile to an affinity in a disparate profile (associated with a disparate user) in order to identify a matching affinity. Once a matching affinity is identified, a message indicating such can be provided to the user along with a request to publish certain revealed information to the disparate user, possibly based upon a mutual exchange of commensurate information from the disparate user.

Correlating private affinities

A software and/or hardware facility for inferring user context and delivering advertisements, such as coupons, using natural language and/or sentiment analysis is disclosed. The facility may infer context information based on a user's emotional state, attitude, needs, or intent from the user's interaction with or through a mobile device. The facility may then determine whether it is appropriate to deliver an advertisement to the user and select an advertisement for delivery. The facility may also determine an appropriate expiration time and/or discount amount for the advertisement.

Context based online advertising

Within a surface computing environment users are provided a seamless and intuitive manner of modifying security levels associated with information. If a modification is to be made the user can perceive the modifications and the result of such modifications, such as on a display. When information is rendered within the surface computing environment and a condition changes, the user can quickly have that information concealed in order to mitigate unauthorized access to the information.

Authentication&#x2014;circles of trust

One or more techniques and/or systems are disclosed that provide for document retrieval where a user can identify key attributes of potential target documents that are desirable (e.g., have a particular semantic content for the user). Further, relevant documents that comprise the desired semantic content can be retrieved. Additionally, the user can provide feedback on the retrieved documents, for example, based on key semantic concepts found in the documents, and the input can be used to update the classification. For example, this process can be iterated to improve the retrieval and precision of documents found through machine learning techniques.

Semantic content searching

A voice message is processed in a distributed system by storing voice message data indicative of a plurality of voice messages on a distributed data store. A distributed data processor accesses the voice messages and extracts desired information from the voice messages. The data processor then augments the data stored in the voice message data store with the extracted information. The user interface component provides user access to the voice messages with the augmented data.

Voice message processing system and method

A method of searching audio data is provided including receiving a query defining multiple phonetic possibilities. The method also includes comparing the query with a lattice of phonetic hypotheses associated with the audio data to identify if at least one of the multiple phonetic possibilities is approximated by at least one phonetic hypothesis in the lattice of phonetic hypotheses.

Phonetic fragment search in speech data

Systems and methods for operating a portable computing device having a broad set of functionality, such as image or audio/video capture capabilities, are provided. One or more specialized modes are provided in a portable computing device, such as a smart phone, for intelligently disabling any distracting functionality of the portable computing device with respect to a target subset of functionality possessed by the device and invoked during operation of a specialized mode. In various embodiments, the special mode for operation of a portable computing device in accordance with the invention is such that functionality that is likely to interfere with a time capture event of the device is disabled or mitigated during operation.

Systems and methods for operating a computing device having image capture capabilities

A computationally efficient and robust pitch detection and tracking system and related methods are presented. According to certain exemplary implementations a method is presented comprising identifying an initial set of pitch period candidates using a first estimation algorithm, filtering the initial set of candidates and passing the filtered candidates through a second, more accurate pitch estimation algorithm to generate a final set of pitch period candidates from which the most likely pitch value is selected.

Speech recognition using dual-pass pitch tracking

Method and apparatus for tracking pitch in audio analysis

In one embodiment, a speech recognition program at a client receives data that is unrecognized, such as an unrecognized word, an unrecognized pronunciation of a known word, an unrecognized dialect of a known, and/or a substantially new word frequency usage. The client transmits the data to a provider, which processes the data into known data, and transmits the known data back to a number of clients, including the client that initially sent the unrecognized data. In one embodiment, the unrecognized data is sent from the client to the provider via a third party, to anonymize the data.

Growing community of inventors

Beijing, China

Eric I-Chao Chang