Samsung Electronics Co., Ltd.

Vijay Srinivasan

Hongxia Jin

Shalinder Singh Sidhu

Abhishek Mukherji

Christian Koehler

Dalkandura Arachchige Kalpa Shashika Silva Gunaratna

Jun Yang

Koichi Mori

Anil Sunder Yadav

Yifei Jiang

Mohammad Mahdi Moazzami

Danny R Bennett

Kiran K Rachuri

Emmanuel M Tapia

Satheesh Sudarsan

Sandeep Nama

Yannick Pellet

Enamul Hoque

Manas Gaur

Emmanuel Munguia-Tapia

Computer-based generation of information-gathering questions in response to a user query can include parsing the user query using natural language processing and extracting from the user query one or more phrases corresponding to a predetermined category. A knowledge database can be accessed, and entities semantically related to each phrase can be extracted therefrom. A query sub-graph representing a relationship between each of the one or more phrases and the entities extracted from the knowledge database can be generated. An expanded user query can be generated by traversing the query sub-graph. Passages semantically related to the expanded user query can be retrieved from one or more passages databases and ranked. A neural question generator can generate a set of information-gathering questions based on the expanded user query and a group of select passages selected from the plurality of passages in accordance with each selected passage's ranking.

Dynamic question generation for information-gathering

Computer-based content understanding can include segmenting an image into a plurality of blocks, wherein each block includes textual information from the image. For each block of the plurality of blocks, encoded feature data is generated by encoding visual information of the block and visual information of one or more neighboring blocks from the plurality of blocks and encoded textual data is generated by encoding the textual information of the block and the textual information of the one or more neighboring blocks. Further, using an entity class prediction model, one or more tokens of the block are classified into one or more entity classes based on a combination of the encoded textual data and the encoded feature data. A plurality of entities can be extracted from the image based on the entity classes of the plurality of blocks.

On-device knowledge extraction from visually rich documents

A method for action automation includes determining, using an electronic device, an action based on domain information. Activity patterns associated with the action are retrieved. For each activity pattern, a candidate action rule is determined. Each candidate action rule specifies one or more pre-conditions when the action occurs. One or more preferred candidate action rules are determined from multiple candidate action rules for automation of the action.

Method for automating actions for an electronic device

An apparatus for contextual execution comprises a processor, and a memory containing instructions, which when executed by the processor, cause the apparatus to receive, from a user terminal, a control input associated with an intent, obtain location data associated with a location of the user terminal, and determine a scored set of execution options associated with the control input. Further, the instructions, when executed by the processor cause the apparatus to obtain a contextual label associated with the location data, the label determined based on the application of one or more adapted pretrained deep learning models to the location data.

System and method for deep labeling

A method includes that for each model from multiple models, evaluating a model prediction accuracy based on a dataset of a user over a first time duration. The dataset includes a sequence of actions with corresponding contexts based on electronic device interactions. Each model is trained to predict a next action at a time point within the first time duration, based on a first behavior sequence over a first time period from the dataset before the time point, a second behavior sequence over a second time period from the dataset before the time point, and context at the time point. A model is selected from the multiple models based on its model prediction accuracy for the user based on a domain. An action to be initiated at a later time using an electronic device of the user is recommended using the selected model during a second time duration.

Predicting user actions on ubiquitous devices

A method comprising receiving different types of crowd-sourced information, the different types of crowd-sourced information relating to a physical store that a plurality of electronic devices have visited. The method further comprises determining a plurality of products available in the physical store based on the different types of crowd-sourced information, and correlating the different types of crowd-sourced information to determine at least one product location of at least one product of the plurality of products available. Each product location of each product identifies a location of the product within the physical store.

Automatic product mapping

Providing sequential behavior-based content may include detecting, using a processor, an event sequence involving at least one device of a user, wherein the event sequence includes a current event and a prior event, correlating, using the processor, the event sequence with a frequent sequential pattern selected from a plurality of frequent sequential patterns associated with the user, and predicting, using the processor, a next event for the user according to a plurality of ordered events specified by the selected frequent sequential pattern. Responsive to predicting the next event, content relating to the next event may be provided to the device using the processor.

Sequential behavior-based content delivery

In one aspect, information of multiple anchor points is received and stored. The information of each anchor point includes Global Positioning System (GPS) data of a particular location and radio frequency (RF) data that was obtained at a device at the particular location. A geo coordinate is determined for an indoor location based on the RF data obtained at the indoor location and the information of the anchor points. Various embodiments pertain to software, systems, devices and methods relating to anchor points and/or the obtaining of a geo coordinate for a location.

Indoor global positioning system

In some embodiments, computer implemented methods, systems, and non-transitory computer readable media determine a first comparison value based on a first comparison between a first sensor signature associated with first set of sensor data of a first device in a first context and a second sensor signature associated with second set of sensor data of a second device. The first comparison is associated with a first authentication type. It is determined whether the first comparison value satisfies a first threshold. It is determined that a user should be authenticated on the second device based on satisfaction of the first threshold.

Systems and methods for device based authentication

A method and device for automatic semantic labeling of unlabeled places using activity recognition. A method includes determining at least one activity based on analyzing electronic device sensor data. Localization for the electronic device is performed to determine location for an unknown semantic place. An observed mapping between the at least one activity and the location for the unknown semantic place is determined. A typical mapping between the at least one activity and at least one semantic place is determined. Using the observed mapping and the typical mapping from one or more other electronic devices, the unknown semantic place is assigned with a semantic place label representing the at least one semantic place for identifying the unknown semantic place. A semantic place map is updated to include the semantic place label.

Automatic semantic labeling based on activity recognition

A method and system for monitoring resource information and user activity. The method includes acquiring one or more data streams from one or more resource meters and one or more electronic device sensors. Discrete events are computed from each data stream. A sequence of discrete sensor-meter event itemsets are extracted based on the events. Frequent sensor-meter event itemsets are discovered from the sequence of discrete event itemsets that occur together, and a frequency of occurrence of each frequent co-occurrence itemset is discovered. Rising sensor-meter event itemsets and falling sensor-meter event itemsets are matched based on appliance state models and the frequency of occurrence of each sensor-meter event itemset. Each individual fixture is identified. Each fixture cluster is classified to a fixture category. Based on the matched fixture events, fixture clusters, and categories, resource usage information and user activities are determined for each fixture usage event identified.

Growing community of inventors

San Jose, CA, United States of America

Vijay Srinivasan