Adobe, Inc.

At+t Corp.

Michael F Kraley

Shawn Alan Gaither

Paul S Kleppner

Trung Huu Bui

Hung Hai Bui

Nikolaos Barmpalios

Pranjal Daga

Scott David Cohen

Brian Lynn Price

Paul John Asente

Walter Wei-Tuh Chang

Randy L Swineford

Andrew Marc Greene

Sangam Pant

Nargol Rezvani

Ruchi Rajiv Deshpande

David M Heath

Dafang He

Walter Chang

Richard J Cohn

R Patrick Johnson

Zoya Bylinskii

Stephen Zagieboylo

Deborah Mendez

Edward Belove

Verena Sabine Kaynig-Fittkau

Sruthi Madapoosi Ravi

Kanchana Sethu

Winslow B Kelley

James A Culbert

Tianyuan Cai

Aleena Niklaus

Techniques are disclosed for readability theme generation. The techniques include obtaining reading formats and generating reading format digital images based on the obtained reading formats. The reading format digital images are encoded using a trained machine learning model as perceptual embeddings. The perceptual embeddings are clustered into reading format clusters and readability themes are determined based on the reading format clusters.

Readability theme generation

Systems and techniques for privacy preserving document analysis are described that derive insights pertaining to a digital document without communication of the content of the digital document. To do so, the privacy preserving document analysis techniques described herein capture visual or contextual features of the digital document and creates a stamp representation that represents these features without included the content of the digital document. The stamp representation is projected into a stamp embedding space based on a stamp encoding model generated through machine learning techniques capturing feature patterns and interaction in the stamp representations. The stamp encoding model exploits these feature interactions to define similarity of source documents based on location within the stamp embedding space. Accordingly, the techniques described herein can determine a similarity of documents without having access to the documents themselves.

Privacy preserving document analysis

The structure of an untagged document can be derived using a predictive model that is trained in a supervised learning framework based on a corpus of tagged training documents. Analyzing the training documents results in a plurality of document part feature vectors, each of which correlates a category defining a document part (for example, 'title' or 'body paragraph') with one or more feature-value pairs (for example, &#x201c;font=Arial&#x201d; or &#x201c;alignment=centered&#x201d;). Any suitable machine learning algorithm can be used to train the predictive model based on the document part feature vectors extracted from the training documents. Once the predictive model has been trained, it can receive feature-value pairs corresponding to a portion of an untagged document and make predictions with respect to the how that document part should be categorized. The predictive model can therefore generate tag metadata that defines a structure of the untagged document in an automated fashion.

Document structure extraction using machine learning

The present invention is directed towards providing automated workflows for the identification of a reading order from text segments extracted from a document. Ordering the text segments is based on trained natural language models. In some embodiments, the workflows are enabled to perform a method for identifying a sequence associated with a portable document. The methods includes iteratively generating a probabilistic language model, receiving the portable document, and selectively extracting features (such as but not limited to text segments) from the document. The method may generate pairs of features (or feature pair from the extracted features). The method may further generate a score for each of the pairs based on the probabilistic language model and determine an order to features based on the scores. The method may provide the extracted features in the determined order.

Probabilistic language models for identifying sequential reading order of discontinuous text segments

The present disclosure relates to systems, methods, and non-transitory computer readable media for generating diverse and realistic synthetic documents using deep learning. In particular, the disclosed systems can utilize a trained neural network to generate realistic image layouts comprising page elements that comply with layout parameters. The disclosed systems can also generate synthetic content corresponding to the page elements within the image layouts. The disclosed systems insert the synthetic content into the corresponding page elements of documents based on the image layouts to generate synthetic documents.

Automatically generating labeled synthetic documents

Automated workflows for identification of reading order from text segments using probabilistic language models

Disclosed systems and methods generate page segmented documents from unstructured vector graphics documents. The page segmentation application executing on a computing device receives as input an unstructured vector graphics document. The application generates an element proposal for each of many areas on a page of the input document tentatively identified as being page elements. The page segmentation application classifies each of the element proposals into one of a plurality of defined type of categories of page elements. The page segmentation application may further refine at least one of the element proposals and select a final element proposal for each element within the unstructured vector document. One or more of the page segmentation steps may be performed using a neural network.

Page segmentation of vector graphics documents

Certain embodiments identify a correct structured reading-order sequence of text segments extracted from a file. A probabilistic language model is generated from a large text corpus to comprise observed word sequence patterns for a given language. The language model measures whether splicing together a first text segment with another continuation text segment results in a phrase that is more likely than a phrase resulting from splicing together the first text segment with other continuation text segments. Sets of text segments, which include a first set with a first text segment and a first continuation text segment as well as a second set with the first text segment and a second continuation text segment, are provided to the probabilistic model. A score indicative of a likelihood of the set providing a correct structured reading-order sequence is obtained for each set of text segments.

Identification of reading order text segments with a probabilistic language model

Disclosed systems and methods generate page segmented documents from unstructured vector graphics documents. The page segmentation application executing on a computing device receives as input an unstructured vector graphics document comprising drawing commands. The application generates an element proposal for each of many areas on a page of the input document tentatively identified as being page elements. Each of the element proposals may be generated at least in part based on the drawing commands. The page segmentation application classifies each of the element proposals into one of a plurality of defined type of categories of page elements at least in part based on the drawing commands. The page segmentation application may further refine at least one of the element proposals and select a final element proposal for each element within the unstructured vector document. One or more of the page segmentation steps may be performed using a neural network.

Techniques are disclosed for efficiently and automatically classifying textual documents or files. In some embodiments, the classification process is integrated into or otherwise made part of the storage function, such that when the user initiates a save process for a given file, the file is processed through a classifier prior to (or contemporaneously with) completing the save function. In some such embodiments, textual content of the file is analyzed using natural language processing to identify a main or substantial concept discussed in the file, and one or more corresponding tags are then assigned to that file. Subsequently, the user can access that file based on the one or more tags, for instance, through a user interface that allows the user to select one or more content categories associated with the assigned tags. The files can be text-based, but may include other content as well, such as images, video, and audio.

Automatic document classification via content analysis at storage time

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, include sending a first electronic document to a first client device over a data network from a web server. The first electronic document includes multiple document elements formatted according to a first document schema for a first client application. The document elements include multiple content elements and one or more presence elements. Presence data is received over the data network at the web server from the first client device. The presence data includes an identification of one or more of the content elements of the first electronic document. The presence data also includes an identification of a status of the first electronic document for the first client device. An updated presence element for the first electronic document, which includes a first presence object, is stored. The first presence object includes the presence data received from the first client device and an identification of the first client device and/or a first user associated with the first client device. The first presence object is sent over the data network from the web server to a second client device.

User presence data for web-based document collaboration

Growing community of inventors

Lexington, MA, United States of America

Michael F Kraley