Palo Alto Research Center Inc.

Evgeniy Bart

Prateek Sarkar

Eric Saund

Robert Roy Price

Juan Liu

John Turner Maxwell, Iii

Shane Patrick Ahern

Michael Roberts

Hoda M A Eldardiry

Christina Pavlopoulou

David Gunning

A computer-implemented system and method for distributed activity detection is provided. Contextual data collected for a user performing an activity is processed on a mobile computing device. The mobile computing device extracts features from the contextual data and compares the features with a set of models. Each model represents an activity. A confidence score is assigned to each model based on the comparison with the features and the mobile computing device transmits the features to a server when the confidence scores for the models are low. The server trains a new model using the features and sends the new model to the mobile computing device.

Computer-implemented system and method for distributed activity detection

Systems and methods described receiving a set of example data and a set of knowledge based data and combine the set of example data and the set of knowledge based data to generate a set of combined data. The combined set can be used to train a machine learning model based on the set of combined data. The machine learning model is applied to a new set of received data for a new subject.

Combined data driven and knowledge driven analytics

One embodiment of the present invention provides a system for multi-domain clustering. During operation, the system collects domain data for at least two domains associated with users, wherein a domain is a source of data describing observable activities of a user. Next, the system estimates a probability distribution for a domain associated with the user. The system also estimates a probability distribution for a second domain associated with the user. Then, the system analyzes the domain data with a multi-domain probability model that includes variables for two or more domains to determine a probability distribution of each domain associated with the probability model and to assign users to clusters associated with user roles.

Detecting anomalies in work practice data by combining multiple domains of information

One embodiment of the present invention provides a system for clustering heterogeneous events. During operation, the system finds a partition of events into clusters such that each cluster includes a set of events. In addition, the system estimates probability distributions for various properties of events associated with each cluster. The system obtains heterogeneous event data, and analyzes the heterogeneous event data to determine the distribution of event properties associated with clusters and to assign events to clusters.

Scalable spatiotemporal clustering of heterogeneous events

A method is provided for parsing a table. The method includes: receiving an input containing the table; finding candidate separators within the table; and determining which candidate separators are at least one of real and spurious by optimizing an objective function over the set of found candidate separators. Suitably, the function measures numerically whether a parse produced by the set of real separators is accurate. The function suitably includes one or more terms that account for multiple aspects of the table including at least two of: quality of candidate separators; coherence of cells within the parse; quality of cells within the parse; coherence of entire rows within the parse; quality of entire rows within the parse; coherence of entire columns within the parse; quality of entire columns within the parse; layout consistency along an axis of the table; and repeatability along the axis of the table.

Parsing tables by probabilistic modeling of perceptual cues

A method is provided for parsing a document having a plurality of lines on which items are listed spanning one or more lines. It includes: obtaining a plurality of candidates, representing hypothetical items within the document, each candidate spanning one or more lines and having a local cost representing a confidence in a quality of the candidate compared to a model; determining labeling costs for intervals of the document defined between pairs of lines, each interval containing candidates therein, each labeling cost reflecting a configuration of the candidates within the interval; identifying a best labeling for each interval based on the labeling costs determined for that interval, the best labeling corresponding to one of the configurations of the candidates within the interval; defining a global objective function; and selecting a subset of the candidates such that the global objective function is optimized, based on the identified best labelings.

Efficient globally optimal interpretation of documents

Methods and system employing the same for optimizing an objective function are provided. The objective function assesses the quality of a candidate solution. One or more variables of an objective function are selected as pivot variables. Each of the variables include one or more candidate values. An upper bound function is generated from the objective function, where the pivot variables are held fixed. For each combination of the candidate values, one or more candidate solutions are searched using the upper bound function. One or more optimal solutions are selected from among the solutions to the searches.

System and method for efficient interpretation of natural images and document images in terms of objects and their parts

Methods and system employing the same for finding repeated structure for data extraction from document images are provided. A reference record and one or more reference fields thereof are identified from a document image. One or more candidate fields are generated for each of the reference fields. One or more best candidate records from the candidate fields are selected using a probabilistic model and an optimal record set is determined from the best candidate records.

Finding repeated structure for data extraction from document images

The present application is a method and system of interpreting an image by finding a configuration of multiple variables which optimizes an objective function with a factorizable upper bound, by applying an iterative algorithm that relies on efficient dynamic ordering of candidate configurations, in a priority queue, in a descending order of an upper bound score. As an example, consider a constellation model for an object. It specifies the appearance models for individual parts of objects, as well as spatial relations among these parts. These are combined into a single function whose value represents the likeness of the object in an image. To find the configuration in which the object is present in the image, we maximize this function over all candidate configurations. The purpose of the iterative algorithm mentioned above is to find such optimal configurations efficiently.

Growing community of inventors

Sunnyvale, CA, United States of America

Evgeniy Bart