Adobe Systems Inc.

Adobe, Inc.

Shawn Alan Gaither

Randy L Swineford

Frederic Thevenet

Michael F Kraley

Nikolaos Barmpalios

Arian Behzadi

Anatole A Matveief

David Parmenter

Jacob Y Stolin

Richard Treitman

Steve Dakin

Kyeung Sub Yeom

James Alisago

Andres Gonzalez

Sachin Soni

Trung Huu Bui

Sanjeev Tagra

Walter Wei-Tuh Chang

Abhishek Jain

Hung Hai Bui

Eylon Stroh

David Rudi Sherry

Andrew Marc Greene

Shagun Kush

Pranjal Daga

Shipali Shetty

Rob K Corell

Nargol Rezvani

Ruchi Rajiv Deshpande

Samarth Gupta

Naqushab Neyazee

Priyank Mathur

Paul John Asente

Richard J Cohn

Mohit Gupta

Ashutosh Mehra

Uttam Dwivedi

Jayant Vaibhav Srivastava

Maurice D Fisher

Verena Sabine Kaynig-Fittkau

Emily Seminerio

Rudi Sherry

Bryan Z Wei

Smitha Bangalore Naresh

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

Techniques are disclosed for identifying asides within a document, and detecting a display order of contents based of the identified asides. In a document, an 'aside' represents a content region of the document that is distinct from the main content regions, and may be visually distinguishable from the main content region. In an example, a document is received, where the document lacks identification of asides. The document is analyzed to identify asides within the document. A display order of contents within the document is then determined, based on the identified asides. For example, in the display order, the asides are ordered between two segments of the main content and/or at a beginning or an end of the main content, but may not be ordered to be embedded in between a segment of the main content. The document is displayed in accordance with the display order.

Asides detection in documents

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

Techniques are provided for identifying structural elements of a document. One Methodology includes generating a first channel of rasterized content by rasterizing a full page of the document and generating one or more additional channels of rasterized content from the page of the document by rasterizing one or more corresponding content types from the page of the document. Each of the one or more additional channels includes a specific type of content that is different from each of the other one or more additional channels. The methodology further includes inputting the first channel of rasterized content and the one or more additional channels of rasterized content into a machine learning (ML) model. The methodology continues with determining location and classification for each of a plurality of structural elements on the page of the document using the ML model.

Multiple channels of rasterized content for page decomposition using machine learning

Techniques described herein implement heading identification and classification for a digital document in a digital medium environment. A document analysis system is leveraged to extract structural features from a digital document, identify heading candidates from among the structural features, validate the headings candidates, and classify validated headings into different headings types. The classified headings are then utilized to generate a sectioned version of the digital document ('sectioned document') that is divided into different sections based on the headings. Further, a document directory is generated that includes the headings and that enables navigation to different sections of the sectioned document.

Heading identification and classification for a digital document

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

Techniques for guiding an interaction with an electronic form via a computing device are described. For example, a form guidance engine is initiated. The form guidance engine maintains a state of the electronic form. The form guidance engine also maintains user information in, for example, a user profile. The user information is associated with a user of the computing device and relates to how the user interacts with the electronic form. Based on the user information, the form guidance engine anticipates a potential interaction with the electronic form given the state. Accordingly and prior to a user request for information about the electronic form, the form guidance engine generates a presentation configured to guide the user to perform the potential interaction with the electronic form via a computing device. A user response to the presentation is received. The form guidance engine performs an action associated with the electronic form and updates the state of the electronic form.

Proactive form guidance for interacting with electronic forms

Techniques are disclosed for identifying and populating static form fields using prior responses to similar form fields. An example method includes receiving an electronic form having, encoded in the form, static text and a static form field for containing information specific to a given user of the electronic form. Using an image recognition algorithm, a fillable form field candidate is identified based on a graphical representation of the static form field. An attribute of the fillable form field candidate can be identified based on the static text. A suggested response for populating the fillable form field candidate is selected from prior responses to other form fields having at least one attribute in common with the identified attribute of the fillable form field candidate. The prior responses are inputs obtained from or associated with the given user. The suggested response is presented to the user for subsequent acceptance or rejection.

Recognition and population of form fields in an electronic document

Techniques for assisting users electronic forms. One exemplary technique involves initiating a workflow assistant based on user input of the user or a predetermined triggering event to prompt the user to complete the electronic form. The workflow assistant downloads or obtains the electronic form based on a user response and identifies a workflow based on the electronic form. The workflow identifies a sequence of user interactions to complete the electronic form. Identifying the workflow includes developing, based on fields in the electronic form, requests for content of fields from the user and possible predicted answer choices for the content of each of the fields. The workflow assistant then executes the workflow to conduct the user interactions to obtain the content for the fields of the electronic form to complete the electronic form.

Interactive electronic form workflow assistant that guides interactions with electronic forms in a conversational manner

Techniques for assisting a user with completing an electronic form via a computing device are described. For example, the user operates the computing device to populate fields of the electronic form. Rather than presenting the electronic form at the computing device in order to populate the fields, a form filling engine reformulates the electronic form into a set of input requests. The form filling engine is local or remote to the computing device. Each input request, when presented, provides guidance about at least one field and solicit a user response that is proper for that field. The form filling engine presents the input requests via the computing device one at a time to solicit the user responses. Data from the user responses are used by the form filling engine to populate the fields of the electronic form.

Form filling engine to populate fields of an electronic form

Systems and methods disclosed herein associate forms with categories based on form features for non-text field characteristics or field-specific text characteristics of the forms. One embodiment provides a method for facilitating searching for a form by associating forms with categories based on form features. The method involves automatically associating, by a processor of a computing device, forms with respective categories based on form features for non-text field characteristics or field-specific text characteristics of the forms and storing the forms and the respective categories associated with the forms at an electronic form search server. Search results are provided from the electronic form search server based on input identifying a search category and a form is identified as a search result based on the form being associated with the search category.

Categorization of forms to aid in form search

Embodiments of the present invention provide systems, methods, and computer storage media directed at predicting values for an electronic form. In embodiments, the method can include forming synonym groupings of form field labels for a number of users. The synonym groupings can be based on an analysis of the similarity of form field values that are associated with form field labels. In embodiments a predictive model may be generated from these synonym groupings. The predictive model can correlate the synonym groupings of one user with synonym groupings of one or more additional users to enable a determination of one or more predicted form field values for the one user based on a queried form field label even though the one user may have never submitted an electronic form with the queried form field label. Other embodiments may be described and/or claimed.

Form value prediction utilizing synonymous field recognition

Growing community of inventors

Raleigh, NC, United States of America

Shawn Alan Gaither