Abstract: This disclosure relates generally to ASR and is particularly directed to automatic, efficient, and intelligent detection of transcription bias in ASR models. Contrary to a tradition approach to the testing of ASR bias, the example implementations disclosed herein do not require actual test speeches and corresponding ground-truth texts. Instead, test speeches may be machine-generated from a pre-constructed reference textual passage according short speech samples of speakers using a neural voice cloning technology. The reference passage may be constructed according to a particular target domain of the ASR model being tested. Bias of the ASR model in various aspects may be identified by analyzing transcribed text from the machine-generated speeches and the reference textual passage. The underlying principles for bias detection may be applied to evaluation of general transcription effectiveness and accuracy of the ASR model.

Abstract: Implementations of the present disclosure include providing a graph representative of a network, a set of nodes representing respective assets, each edge representing one or more lateral paths between assets, the graph data including configurations affecting at least one impact that has an effect on an asset, determining multiple sets of fixes for configurations, each fix having a cost associated therewith, incorporating fix data of the sets of fixes into the graph, defining a set of fixes including one or more fixes from the multiple sets of fixes by defining an optimization problem that identifies one or more impacts that are to be nullified and executing resolving the optimization problem to define the set of fixes, each fix in the set of fixes being associated with a respective configuration in the graph, and scheduling performance of each fix in the set of fixes based on one or more operational constraints.

Abstract: This document described pressure-sensitive human machine interface (HMI) devices. In some aspects, a pressure-sensitive human-machine interface (HMI) device may include a surface element comprising an outer surface for receiving pressure inputs. The HMI device may also include multiple pressure-sensitive input regions that each include one or more positive electrodes and one or more return electrodes. The HMI device may also include a pressure-sensitive conductive sheet disposed adjacent to or between the one or more positive electrodes and one or more return electrodes of the multiple pressure-sensitive input regions.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer-storage media, for dynamic user data filtering. In some implementations, a method includes determining one or more values using data representing a sequence of one or more types of interactions between a user and content; using the one or more values to determine whether to include each interaction of the one or more types of interactions in the sequence within a reduced user data set; generating the reduced user data set by removing one or more interactions from the sequence based on determining not to include the one or more interactions using the one or more values; and providing the reduced user data set to a processing server.

Abstract: In some implementations, a device may receive, from a user device, a query for information from a user. The device may receive, from an enterprise system, data identifying products and services to be offered by the user. The device may process the query and the data, with a model, to identify a subject of the query and a context for the subject. The device may process the subject and the context, with a model, to determine a workflow for the user. The device may process the data, the subject of the query, and the context for the subject, with a model, to determine a key performance indicator or a decision associated with one of the products or the services. The device may provide information identifying the workflow, the key performance indicator, and/or the decision to the user device via a virtual assistant or a user interface.

Abstract: In some implementations, a device may generate, based at least in part on a first set of inputs, a serverless software development environment associated with a set of cloud resources. The device may generate, based at least in part on a first machine learning model, a technology stack recommendation having a set of associated tools for performing a software development task. The device may instantiate the selected technology stack in the serverless software development environment and generate a set of applications based at least in part on executing the set of tools. The device may deploy the set of applications in one or more serverless application environments. The device may use machine learning to observe deployed applications, detect hidden anomalies, and perform root-cause analysis, thereby providing a lean and sustainable serverless environment.

Abstract: A device may receive project management data associated with development of a software product and may process a first portion of the project management data, with first models, to generate timeliness scores and an overall timeliness score for the software product. The device may process a second portion of the project management data, with second models, to generate quality scores and an overall quality score for the software product and may process a third portion of the project management data, with third models, to generate product readiness scores and an overall product readiness score for the software product. The device may utilize a fourth machine learning model, with the overall timeliness score, the overall quality score, and the overall product readiness score, to generate a success probability for the software product and may perform one or more actions based on the success probability for the software product.

Abstract: A device may receive security data identifying assets of an entity, security issues associated with the assets, and objectives associated with the assets and may utilize a data model to generate, based on the security data, asset related data identifying mapped sets of security data. The device may process a first portion of the asset related data, with a first model, to calculate an asset risk likelihood score for an asset of the assets and may process a second portion of the asset related data, with a second model, to calculate an asset criticality score for the asset. The device may process a third portion of the asset related data, with a third model, to calculate an asset control effectiveness score for the asset and may combine the scores to generate a security risk score for the asset. The device may provide the security risk score for display.

Abstract: Methods, systems, and apparatus for solving computational tasks using quantum computing resources. In one aspect a method includes receiving, at a quantum formulation solver, data representing a computational task to be performed; deriving, by the quantum formulation solver, a formulation of the data representing the computational task that is formulated for a selected type of quantum computing resource; routing, by the quantum formulation solver, the formulation of the data representing the computational task to a quantum computing resource of the selected type to obtain data representing a solution to the computational task; generating, at the quantum formulation solver, output data including data representing a solution to the computational task; and receiving, at a broker, the output data and generating one or more actions to be taken based on the output data.

Abstract: In some examples, scalable source code vulnerability remediation may include receiving source code that includes at least one vulnerability, and receiving remediated code that remediates the at least one vulnerability associated with the source code. At least one machine learning model may be trained to analyze a vulnerable code snippet of the source code. The vulnerable code snippet may correspond to the at least one vulnerability associated with the source code. The machine learning model may be trained to generate, for the vulnerable code snippet, a remediated code snippet to remediate the at least one vulnerability associated with the source code. The remediated code snippet may be validated based on an analysis of whether the remediated code snippet remediates the at least one vulnerability associated with the source code.

Abstract: Implementations of the present disclosure include receiving analytical attack graph data representative of an analytical attack graph, the analytical attack graph including: one or more rule nodes each representing a network configuration rule; and one or more impact nodes each representing an impact of one or more respective network configuration rules; converting the analytical attack graph to a tactic graph including one or more tactic nodes, each tactic node representing at least one rule node and at least one impact node; determining one or more paths of the tactic graph that lead to a particular network impact; generating a process model based on the paths that lead to the particular network impact, the process model representing network activity for execution of a process that leads to the particular network impact; and executing one or more remedial actions based on the process model to mitigate cyber-security risk to the enterprise network.

Abstract: A device may receive and transform metric data and share of voice data, associated with digital marketing by an entity, into transformed data, may generate model data from the transformed data, and may divide the model data into training data, test data, and validation data. The device may train models, with the training data, to generate training results, and may process the test data, with the models, to generate test results. The device may process the validation data, with the models, to generate validation results, and may select a first model, a second model, and a third model based on the results. The device may utilize the first model to predict a share of voice, and may utilize the second model to predict a click through rate. The device may utilize the third model to predict a conversion rate, and may perform actions based on the predicted data.

Abstract: Automated response generation systems and methods are disclosed. The systems can include a deep learning model specially configured to apply inferencing techniques to redesign natural language querying systems for use over knowledge graphs. The disclosed systems and methods provide a model for inferencing referred to as a Hierarchical Recurrent Path Encoder (HRPE). An entity extraction and linking module as well as a data conversion and generation module process the content of a given query. The output is processed by the proposed model to generate inferred answers.

Abstract: Implementations are directed to configuring a set of applications and one or more modules associated with each application, wherein the one or more modules of an application comprise functional components that are bundled into the application, wherein each application is associated with a site of an on-premise system where the application is to be deployed; creating a process flow that includes a plurality of nodes, each node corresponding to a process executed at the site; associating a collection of applications to each node included in the process flow, wherein the collection of applications are selected from the set of applications, wherein the set of applications are categorized based on a relevance score of each application; and deploying the process flow and the collection of applications associated with each node to corresponding on-premise edge devices of the on-premise system based on the site of each application.

Abstract: A mobile unit that includes a compressed gas canister, a brake hose coupled to the compressed gas canister, a radar, and a processor and a storage device. The storage device stores instructions that are operable, when executed by the processor, to cause the processor to perform operations of obtaining radar data from the radar, determining, based on the radar data, a distance of the mobile unit to an object, determining that the distance of the mobile unit to the object satisfies a braking criteria, and based on determining that the distance of the mobile unit to the object satisfies the braking criteria, releasing gas from the compressed gas canister through the brake hose coupled to the compressed gas canister and into the brake pipe of a train car.

Abstract: Springs can provide energy return and have a conductivity that changes in relation to an amount of strain or deformation of the spring. In some embodiments, the springs are made by multi-material 3D printing (additive manufacturing). Such springs made by multi-material 3D printing may include a first material that is electrically non-conductive and a second material that electrically conductive. The extent of deformation or strain of the spring may be determined or estimated by measuring the conductivity or resistivity of the electrically conductive material portion of the spring.

Abstract: Methods, systems, and computer-readable storage media for receiving data representative of a physical entity, generating an initial knowledge graph representative of a process that is executed by the physical entity based on the data, enriching the initial knowledge graph to provide a process aware energy consumption (PAEC) digital twin of the process as an enriched knowledge graph, providing at least two permutations based on the PAEC digital twin, executing analytics at least partially based on the at least two permutations to provide one or more recommendations, and executing at least one recommendation to optimize energy consumption of the physical entity.

Abstract: Methods, systems, and computer-readable storage media for receiving a AAG from computer-readable memory, generating from logical network ontology data, asset inventory data, and asset communication data, a logical topology of the enterprise network as a computer-readable data structure, defining, at least partially by executing community detection over the logical topology, a sub-set of groups within the enterprise network, each group representing a process of a plurality of process, each process being at least partially executed by one or more assets within the enterprise network, processing the AAG based on the sub-set of groups and data from one or more contextual data sources to provide the process aware AAG, the process aware AAG defining a mapping between an infrastructure-layer of the enterprise network and a process-layer of the enterprise network, and executing one or more remedial actions in the enterprise network in response to analytics executed on the process aware AAG.

Abstract: Methods, systems, and computer-readable storage media for receiving, by an aromatic simulation platform, a recipe including descriptions indicative of ingredients of a consumable, processing the recipe through a first neural network to provide a recipe embedding, processing an ingredients profile to determine an aroma compounds profile representing the ingredients of the ingredients profile, processing the aroma compounds profile through a second neural network to provide an aroma embedding, and processing the recipe embedding and the aroma embedding through a third neural network to provide an aroma profile representative of the consumable of the recipe.

Abstract: Systems and methods are provided to identify security vulnerabilities related to containerization platforms. Container images may be received from a repository, and scanned for security vulnerabilities. Containers may be automatically generated and updated with security updates when the images are extracted and identified. Updated versions of images may be generated based on the updated containers. Stored security vulnerability may be automatically updated with CVE information received from external databases at regular intervals, or upon receiving a scan request. Scan results may be generated, stored and compared. Vulnerability comparisons may be generated for an initial version of an image and an updated version of the image that includes the implemented security updates that rectify the identifiable security vulnerabilities.