Microsoft Technology Licensing, LLC

Vishal Chhabilbhai Thakkar

Yan Wang

Arun Sacheti

Surendra Ulabala

David J Simons

Houdong Hu

Robert Osborne

Ajay P Chitturi

Girish Premchandran

Rakesh Tangirala

Ye Wu

Leon Chih Wen Wong

Ling-Cheung Aaron Lo

Shloak Jain

A computer-implemented technique generates a dense embedding vector that provides a distributed representation of input text. The technique includes: generating an input term-frequency (TF) vector of dimension g that includes frequency information relating to frequency of occurrence of terms in an instance of input text; using a TF-modifying component to modify the term-specific frequency information in the input TF vector by respective machine-trained weighting factors, to produce an intermediate vector of dimension g; using a projection component to project the intermediate vector of dimension g into an embedding vector of dimension k, where k is less than g. Both the TF-modifying component and the projection component use respective machine-trained neural networks. An application performs any of a retrieval-based function, a recognition-based function, a recommendation-based function, a classification-based function, etc. based on the embedding vector.

Encoder using machine-trained term frequency weighting factors that produces a dense embedding vector

Representative embodiments disclose mechanisms to create a text stream from raw OCR outputs. The raw OCR output comprises a plurality of bounding boxes, each bounding box defining a region containing text which has been recognized by the OCR system. A weight matrix is calculated that comprises a weight for each pair of bounding boxes. The weight representing the probability that a pair of bounding boxes belongs to the same cluster. The bounding boxes are then clustered along the weights. The resulting clusters are first ordered using an ordering criteria. The bounding boxes within each cluster are then ordered according to a second ordering criteria. The ordered clusters and bounding boxes are then arranged into a text stream.

Structural clustering and alignment of OCR results

A system and methods for facilitating multiparty communications sessions with a plurality of participants and for dynamically designating a communications session manager. More particularly, a system and methods, including protocols, for: establishing a multiparty communications session between a plurality of participants and their respective communication devices in a full-mesh topology; establishing a dynamically designated session manager uniquely associated with a first participant of the multiparty session and, hence, such participant's communication device; as necessary, adding and removing participants and their respective communication devices from the multiparty communications session; and, designating, or electing, another participant as session manager when the first participant exits the multiparty communications session. The methods also enable participants to maintain consistent lists of session participants and establish peer to peer communication links between the participants.

System and methods for facilitating a multiparty communications session with a dynamically designated session manager

A method for navigating between elements on a page of content using an input in a handheld device. The method includes receiving an input corresponding to a navigation direction, determining an element that is radially closest to a current focus element in the navigation direction, and setting the element as the next current focus element. The input may be received from a four-way directional pad associated with the handheld device.

Method for navigation between elements on a page of content in a handheld device

Growing community of inventors

Kirkland, WA, United States of America

Vishal Chhabilbhai Thakkar