Patents.

Method and system for predicting task completion of a time period based on task completion rates of prior time periods using machine learning

Multi-tasker

The disclosure is directed to presenting a unified view of tasks and applications across multiple computing devices of a user. A unified view of the tasks can be presented using a task manager. The task manager presents a list of tasks that are executing at each of the devices. The user can perform multiple actions on the tasks, e.g., launch on a first device a task executing on a second device from the first device, or terminate from the first device a task that is executing on the second device. A unified view of the applications can be presented using an app tray. The app tray displays a list of apps installed on each of the devices of the user. In displaying the list on a particular device, the app tray displays the apps that are not installed on the particular device in a distinct format, e.g., as shadow icons.

Splitting transaction and analysis queries

Secure storage audit verification system

A system for secure storage audit verification includes a transaction pool and a processor. The transaction pool stores transactions records for a Key Release System (KRS) Audit Agent or a Key Management System (KMS) Audit Agent. The processor is configured to select a transaction, wherein the transaction is awaiting verification in the transaction pool; determine whether the transaction is old; in response to the transaction not being old, select a new transaction; and in response to the transaction being old, indicate to suspend processing of the transaction.

Systems and methods for a multi-core optimized recurrent neural network

Systems and methods for implementing recurrent neural networks optimized for multi-core processors by keeping recurrent weights resident on chip — the patent counterpart of the Persistent RNNs work.

System and method for communicating through multiple endpoints

Compacting data history files

Methods, systems, and apparatus for obtaining one or more metadata files, determining, by one or more computers and in accordance with a minor compaction setting, to perform a minor compaction of the one or more metadata files, creating one or more intermediate metadata files that each include at least compacted contacts of one or more of the metadata files, according to the determination to perform minor compaction of the one or more metadata files, determining, in accordance with a major compaction setting, to perform a major compaction of one or more of the intermediate metadata files, and creating one or more snapshot metadata files that each include at least compacted contents of one or more of the intermediate metadata files, according to the determination to perform the major compaction of one or more of the intermediate metadata files.

Recognition of biases in data and models

A method of training a predictive model to predict a likely field value for one or more user selected fields within an application, including analyzing a pre-existing user-provided data set for biases.

Decomposition of queries

Resource and metric ranking by differential analysis

Systems and methods for micro-batch processing of data

Brain wave processing for diagnosis of a subject

Embodiments relate to digital image processing for diagnosis of a subject. More specifically, the embodiments relate to automation of diagnoses through data interpretation. An image is acquired from the subject.

Autonomous vehicle platform and safety architecture

In embodiments of an autonomous vehicle platform and safety architecture, safety managers of a safety-critical system monitor outputs of linked components of the safety-critical system. The linked components comprise at least three components, each of which is configured to produce output indicative of a same event independent from the other linked components by using different input information than the other linked components. The safety managers also compare the outputs of the linked components to determine whether each output indicates the occurrence of a same event. When the output of one linked component does not indicate the occurrence of an event that is indicated by the outputs of the other linked components, the safety managers identify the one linked component as having failed. Based on this, the outputs of the other linked components are used to carry out operations of the safety-critical system without using the output of the failed component.

Browser-based payment for content

Disclosed are various embodiments for browser-based payment for content. A first request for content is sent to a network content server. A response protocol header is received indicating that a payment is sought for the content in response to the first request. Payment manager code integral to a browser or a browser plug-in is executed in response to receiving the response protocol header. A payment-signifying token is received from a payment provider in response to consummating the payment. A second request for the content is sent to the network content server, where the second request includes a request protocol header specifying the payment-signifying token.

Detecting cracks on operatively coupled touchscreen display using pressure sense

A method of identifying a crack in a touchscreen display system includes measuring a distance between a first layer of the touchscreen display system and a second layer of the touchscreen display system, detecting a change in the distance at a portion of the touchscreen display system, identifying a crack in the first layer at the portion of the touchscreen display system based on the change in the distance, and adjusting one or more parameters of the touchscreen display system based on the crack in the first layer.

Deep learning models for speech recognition

Embodiments of state-of-the-art speech recognition systems developed using end-to-end deep learning (the Deep Speech line of work), replacing hand-engineered pipeline components with neural networks trained on large corpora.

Reducing instances of inclusion of data associated with hindsight bias in a training set of data for a machine learning system

Instances of data associated with hindsight bias in a machine learning training set can be reduced: a first set of fields is analysed against a second field corresponding to the event to be predicted, and leaking fields are removed. (European application; family also filed as WO, US, CN, JP)

Detecting malicious domains and client addresses in DNS traffic

IoT device risk assessment

Systems and methods for automatic unit selection and target decomposition for sequence labelling

Described herein are systems and methods for automatic unit selection and target decomposition for sequence labelling. Embodiments include a new loss function called Gram-Connectionist Temporal Classification (CTC) loss that extend the popular CTC loss function criterion to alleviate prior limitations. While preserving the advantages of CTC, Gram-CTC automatically learns the best set of basic units (grams), as well as the most suitable decomposition of target sequences. Unlike CTC, embodiments of Gram-CTC allow a model to output variable number of characters at each time step, which enables the model to capture longer term dependency and improves the computational efficiency. It is also demonstrated that embodiments of Gram-CTC improve CTC in terms of both performance and efficiency on the large vocabulary speech recognition task at multiple scales of data, and that systems that employ an embodiment of Gram-CTC can outperform the state-of-the-art on a standard speech benchmark.

Methods and system for fast, adaptive correction of misspells

Embodiments are directed to a spellcheck module for an enterprise search engine. The spellcheck module includes a candidate suggestion generation module that generates a number of candidate words that may be the correction of the misspelled word.

Systems and methods of monitoring a network topology

Method and device for obtaining a proof of work in a computer network

Cluster Instance Management System

A system for cluster management comprises a status monitor and an instance replacement manager. The status monitor is for monitoring status of an instance of a set of instances on a cluster provider. The instance replacement manager is for determining a replacement strategy for the instance in the event the instance does not respond. The replacement strategy for the instance is based at least in part on a management criteria for on-demand instances and spot instances on the cluster provider.

Medical scan comparison system

A medical scan comparison system is operable to receive a medical scan via a network and to generate similar scan data. The similar scan data includes a subset of medical scans from a medical scan database that are determined to be similar to the received scan.

Measuring utilization of resources in datacenters

For measuring component utilization in a computing system, a server energy utilization reading of a statistical significant number of servers out of a total number of servers located in the datacenter is obtained by measuring, at predetermined intervals, a collective energy consumed by all processing components within each server. The collective energy is measured by virtually probing thereby monitoring an energy consumption of individual ones of all the processing components to each collect an individual energy utilization reading, where the individual energy utilization reading is aggregated over a predetermined time period to collect an energy consumption pattern associated with the server utilization reading.

Hot-cold data separation method for reducing write amplification in key-value stores

A method incorporating hot-cold data separation during compaction in data storage systems: data is classified during compaction into two groups by update frequency, and the colder group is rewritten into a lower storage level that is compacted less frequently, reducing write amplification in key-value stores.

Selective growth of seed layer for magneto-resistive random access memory

Skill extraction system

In an example, disclosed is a machine automated method of identifying a set of skills. In some examples, the method includes extracting a plurality of skill seed phrases from a plurality of member profiles of a social networking site, creating a plurality of disambiguated skill seed phrases by disambiguating the plurality of skill seed phrases using one or more computer processors, and de-duplicating the plurality of disambiguated skill seed phrases to create a plurality of de-duplicated skill seed phrases.

Extraction of system administrator actions to a workflow providing a resolution to a system issue

A method for creating a workflow using system administrator actions to resolve a system issue is provided. The method may include assigning a ticket to a category. The method may include capturing a first snapshot of a first system state of a machine before a system administrator begins a corrective action. The method may include capturing a second snapshot of a second system state of the machine after system administrator corrective action. The method may include creating a difference set based on a comparison of the first snapshot and the second snapshot. The method may include generating summaries of administrator action. The method may include dividing the summaries into groups of summaries. The method may include constructing a candidate workflow for each group. The method may include presenting the candidate workflow to a subject matter expert. The method may include storing the reviewed candidate workflow in a repository.

Multiversion concurrency control of database records with uncommitted transactions

Medical scan image analysis system

A medical scan image analysis system is operable to receive a plurality of medical scans that represent a three-dimensional anatomical region and include a plurality of cross-sectional images. A computer vision model is trained on the plurality of medical scans and used to generate inference data for new medical scans.

Method for synchronization of repository data using data criteria

In an embodiment, a source repository system sends an ending timestamp to determine the range of data that will be obtained from the source repository system. In an embodiment, data from the source repository system is grouped in clusters of data items for transferring the data to a target system.

End-to-end speech recognition

Embodiments of end-to-end deep learning systems for speech recognition (Deep Speech 2) that handle a diverse variety of speech including noisy environments, accents and different languages, using recurrent/convolutional neural networks trained with CTC.

Systems and methods for operating a server-side data abstraction layer

A method receives a first request from a client object at a device specifying a data source, uploads the data, stores it as a plurality of first columns, and instantiates a server object providing access to those columns; a second request specifies a transformation and yields additional typed columns.

Determining relevancy of discussion topics

Embodiments are directed to a computer implemented method of determining the relevancy of multiple sources. The method includes receiving a selection of one or more sources for monitoring. Thereafter, a relevancy factor can be assigned to each source. Data from one or more targets existing on each source can be gathered. In addition, a decay factor can be applied to the data to weight the data based on how recent the data is. The relevancy factor and the decay factor can be used to determine a relevancy score to determine the relevancy of the data. Other embodiments are also disclosed.

Pattern match-based detection in IoT security

Feature-based classification of individual domain queries

System and method for group control in a metaverse application

A metaverse system includes a client computer coupled to a network, a metaverse server coupled to the client computer, the metaverse server to host a metaverse application, and a group link engine coupled to the server. The engine opens a group link connection in response to a request from a user. A linked group associated with the connection includes at least two linked users, and at least one of the linked users is designated as a leader, and each of the at least two linked users is associated with an avatar different from the avatar associated with the other. The engine includes a group link controller to allow the leader to control an action of a second avatar of a second user. The action of the second avatar can be viewed separately from each avatar associated with the at least one linked users that are designated as the leader.

Metadata tag description generation

Method and system for distributed processing in a messaging platform

System and Method for Predicting Organizational Outcomes

A system for modeling an entity includes a prediction module configured to identify one of a plurality of entities using a set of criteria filters to construct a model for the entity with model features built from data about the entity. The prediction module computes a diffusion model coefficient based on a diffusion parameter vector, and a drift model coefficient based on a drift parameter vector. The module computes a predicted entity success probability and a success confidence interval. A portfolio selection module receives a score measuring success for the plurality of entities based on the model, orders the scores in rank order, and forms a portfolio from the top n scoring entities.

Iteratively updating parameters for dialog states

Lithium sulfur cell with dopant

Mobile device accident avoidance system

A system, method and program product for providing accident avoidance for a mobile device having a camera integrated therein. A system is disclosed that includes: a camera orientation device that is mountable on a mobile device and includes a mechanism to reorient a line of sight of the camera in ...

Anomaly detection in distributed ledger systems

Streaming decision in the cloud

Generating streaming decisions at one or more remote servers on behalf of a local client is disclosed. An indication of a content request, multi-dimensional client information, and local state information are received from a client device. A record specific to the client device is generated, in a data store, using the received local state information. The data store includes a plurality of records corresponding to respective different client devices. A set of instructions usable by the client device to obtain and play content is determined based at least in part the local state information received from the client device. The determined set of instructions is transmitted to the client device.

Comprehensive safety schema for ensuring appropriateness of language in online chat

A method is disclosed for evaluating a chat message sent between users of an online environment. The method may include associating each word in the chat message with metadata. The metadata identifies a word type and usage for each word in the chat message.

Re-ranking search results for location refining and diversity

Translation of natural language questions and requests to a structured query format

A natural language query (NLQ) is translated to a structured data query (e.g., a SQL statement) by extracting entities from the NLQ and replacing them with generic variables to form a generic query.