公开(公告)号：US20180329978A1

公开(公告)日：2018-11-15

申请号：US15863337

申请日：2018-01-05

Applicant: Microsoft Technology Licensing, LLC

Inventor： Gary J. Sullivan ,  Leon Bottou ,  Sandeep Kanumuri ,  Yongjun Wu

IPC: G06F17/30 ,  H04N19/44

CPC classification number: G06F17/30598 ,  H04N19/44

Abstract: Innovations for category-prefixed data batching (“CPDB”) of entropy-coded data or other payload data for coded media data, as well as innovations for corresponding recovery of the entropy-coded data (or other payload data) formatted with CPDB. The CPDB can be used in conjunction with coding/decoding for video content, image content, audio content or another type of content. For example, after receiving coded media data in multiple categories from encoding units, a formatting tool formats payload data with CPDB, generating a batch prefix for a batch of the CPDB-formatted payload data. The batch prefix includes a category identifier and a data quantity indicator. The formatting tool outputs the CPDB-formatted payload data to a bitstream. At the decoder side, a formatting tool receives the CPDB-formatted payload data in a bitstream, recovers the payload data from the CPDB-formatted payload data, and outputs the payload data (e.g., to decoding units).

公开(公告)号：US09779081B2

公开(公告)日：2017-10-03

申请号：US15135266

申请日：2016-04-21

Applicant: MICROSOFT TECHNOLOGY LICENSING, LLC

Inventor： Patrice Y. Simard ,  David Max Chickering ,  David G. Grangier ,  Denis X. Charles ,  Leon Bottou ,  Carlos Garcia Jurado Suarez

IPC: G06F17/27 ,  H04L1/00 ,  G06N99/00 ,  G06N7/00 ,  G06F3/0482 ,  G06F17/30

CPC classification number: G06F17/2735 ,  G06F3/0482 ,  G06F17/2785 ,  G06F17/30864 ,  G06N7/005 ,  G06N99/005 ,  H04L1/0072 ,  H04L1/0079

Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.