Abstract: Methods and systems of translation include performing a search of translation hypotheses for input source code in a first programming language to a second programming language. A constraint is extracted from the input source code. The constraint is applied to the translation hypotheses to generate a source code output in the second programming language.

Abstract: Test cases are generated satisfying constraints from natural language descriptions in code by parsing source code to extract variables, and extracting constraints from a natural language description in the source code to retrieve boundary conditions on the variables. Mapping between the variables extracted from the source code and the constraints from the natural language description; and generating input variables satisfying the constraints provided from the natural language description. The method may further include executing the source code using the input variables satisfying the constraints provided from the natural language description.

Abstract: Disclosed are techniques for wireless communication. In an aspect, an asset tracking device receives a set of group scheduling parameters for a plurality of asset tracking devices assigned to a plurality of assets in a shipment, wherein the set of group scheduling parameters comprises a global wakeup start time and a time interval between consecutive wakeup times, wherein the shipment comprises a plurality of stops, including a starting stop, one or more intermediate stops, and an ending stop, and wherein each of the plurality of asset tracking devices has a target stop of the plurality of stops, and, at each wakeup time of at least a set of wakeup times of the consecutive wakeup times: obtains one or more positioning measurements, transmits the one or more positioning measurements, synchronizes a local clock of the asset tracking device to a global time protocol, and transitions to a sleep mode.

Abstract: A system that automatically reduces the time to execute software testing through intelligent test selection and execution. The system automatically detects what tests to execute based on code that has been changed, which is a subset of the entire list of tests to run for the block of code. Once the subset of tests is identified, annotations for tests are processed to update the subset as desired by the code administrator. Once updated, the system then automatically obtains the tests for the updated subset of tests. The tests to be executed are then distributed into groups or buckets. The distribution is set so that each group of tests will have as close to the same execution time as possible. The tests in each group or bucket are then executed concurrently with the other grouped tests.

Abstract: A modular two-stage neural architecture is used in translating a natural language question into a logic form such as a SPARQL Protocol and RDF Query Language (SPARQL) query. In a first stage, a neural machine translation (NMT)-based sequence-to-sequence (Seq2Seq) model translates a question into a sketch of the desired SPARQL query called a SPARQL silhouette. In a second stage a neural graph search module predicts the correct relations in the underlying knowledge graph.

Abstract: An embodiment includes decomposing a natural language assertion into a natural language question and answer pair that includes an initial question and an initial answer. The embodiment translates the initial question into a structured knowledge graph query and then performs an iterative process comprising iterative querying of a knowledge graph and evaluating of corresponding query responses resulting in respective confidence scores. A first iteration of the iterative process comprises querying of the knowledge graph to retrieve a first predicted answer, then determining whether a degree of similarity between the initial answer and the first predicted answer meets a threshold criterion. If not, the first predicted query is altered and used for querying the knowledge graph in a subsequent iteration of the iterative process. The embodiment also generates an assertion correctness score indicative of a degree of confidence that the assertion is factual using the respective confidence scores.

Abstract: Methods, systems, and computer program products for zero-shot entity linking based on symbolic information are provided herein. A computer-implemented method includes obtaining a knowledge graph comprising a set of entities and a training dataset comprising text samples for at least a subset of the entities in the knowledge graph; training a machine learning model to map an entity mention substring of a given sample of text to one corresponding entity in the set of entities, wherein the machine learning model is trained using a multi-task machine learning framework using symbolic information extracted from the knowledge graph; and mapping an entity mention substring of a new sample of text to one of the entities in the set using the trained machine learning model.

Abstract: Implementations of the subject technology provide systems and methods for multi-mode voice triggering for audio devices. An audio device may store multiple voice recognition models, each trained to detect a single corresponding trigger phrase. So that the audio device can detect a specific one of the multiple trigger phrases without consuming the processing and/or power resources to run a voice recognition model that can differentiate between different trigger phrases, the audio device pre-loads a selected one of the voice recognition models for an expected trigger phrase into a processor of the audio device. The audio device may select the one of the voice recognition models for the expected trigger phrase based on a type of a companion device that is communicatively coupled to the audio device.

Abstract: One or more systems, devices, computer program products and/or computer-implemented methods of use provided herein relate to a process for matching a word subset to a database entity. A system can comprise a memory that stores computer executable components, and a processor that executes the computer executable components stored in the memory, wherein the computer executable components can comprise an identification component that outputs a word subset based on a word-based input file, and a mapping component that, based on a rules-based process employing soft matching, maps the word subset to a category comprising a value for being correlated to the word subset. The rules-based process employed by the mapping component can comprise word vector matching or fuzzy string matching.

Abstract: Implementations of the subject technology provide systems and methods for multi-mode voice triggering for audio devices. An audio device may store multiple voice recognition models, each trained to detect a single corresponding trigger phrase. So that the audio device can detect a specific one of the multiple trigger phrases without consuming the processing and/or power resources to run a voice recognition model that can differentiate between different trigger phrases, the audio device pre-loads a selected one of the voice recognition models for an expected trigger phrase into a processor of the audio device. The audio device may select the one of the voice recognition models for the expected trigger phrase based on a type of a companion device that is communicatively coupled to the audio device.

Abstract: One or more computer processors parse a received natural language question into an abstract meaning representation (AMR) graph. The one or more computer processors enrich the AMR graph into an extended AMR graph. The one or more computer processors transform the extended AMR graph into a query graph utilizing a path-based approach, wherein the query graph is a directed edge-labeled graph. The one or more computer processors generate one or more answers to the natural language question through one or more queries created utilizing the query graph.

Abstract: A modular two-stage neural architecture is used in translating a natural language question into a logic form such as a SPARQL Protocol and RDF Query Language (SPARQL) query. In a first stage, a neural machine translation (NMT)-based sequence-to-sequence (Seq2Seq) model translates a question into a sketch of the desired SPARQL query called a SPARQL silhouette. In a second stage a neural graph search module predicts the correct relations in the underlying knowledge graph.

Abstract: An embodiment includes decomposing a natural language assertion into a natural language question and answer pair that includes an initial question and an initial answer. The embodiment translates the initial question into a structured knowledge graph query and then performs an iterative process comprising iterative querying of a knowledge graph and evaluating of corresponding query responses resulting in respective confidence scores. A first iteration of the iterative process comprises querying of the knowledge graph to retrieve a first predicted answer, then determining whether a degree of similarity between the initial answer and the first predicted answer meets a threshold criterion. If not, the first predicted query is altered and used for querying the knowledge graph in a subsequent iteration of the iterative process. The embodiment also generates an assertion correctness score indicative of a degree of confidence that the assertion is factual using the respective confidence scores.

Abstract: According to one embodiment of the present invention, a system transforms an application for a distributed computing environment. The system comprises one or more memories, and at least one processor coupled to the one or more memories. The system analyzes a description of user intent to extract information for transforming the application. The extracted information indicates functionalities for the distributed computing environment. A plurality of software artifacts of the application are mapped to the functionalities. The plurality of software artifacts form different groups of software artifacts. Remaining software artifacts of the application are assigned into the different groups based on a remaining software artifact corresponding to a mapped software artifact of a group. The different groups correspond to microservices for the distributed computing environment. The microservices for the distributed computing environment are presented based on the different groups.

Abstract: Implementations of the subject technology provide systems and methods for multi-mode voice triggering for audio devices. An audio device may store multiple voice recognition models, each trained to detect a single corresponding trigger phrase. So that the audio device can detect a specific one of the multiple trigger phrases without consuming the processing and/or power resources to run a voice recognition model that can differentiate between different trigger phrases, the audio device pre-loads a selected one of the voice recognition models for an expected trigger phrase into a processor of the audio device. The audio device may select the one of the voice recognition models for the expected trigger phrase based on a type of a companion device that is communicatively coupled to the audio device.

Abstract: Methods, systems, and computer program products for zero-shot entity linking based on symbolic information are provided herein. A computer-implemented method includes obtaining a knowledge graph comprising a set of entities and a training dataset comprising text samples for at least a subset of the entities in the knowledge graph; training a machine learning model to map an entity mention substring of a given sample of text to one corresponding entity in the set of entities, wherein the machine learning model is trained using a multi-task machine learning framework using symbolic information extracted from the knowledge graph; and mapping an entity mention substring of a new sample of text to one of the entities in the set using the trained machine learning model.

Abstract: Implementations of the subject technology provide systems and methods for multi-mode voice triggering for audio devices. An audio device may store multiple voice recognition models, each trained to detect a single corresponding trigger phrase. So that the audio device can detect a specific one of the multiple trigger phrases without consuming the processing and/or power resources to run a voice recognition model that can differentiate between different trigger phrases, the audio device pre-loads a selected one of the voice recognition models for an expected trigger phrase into a processor of the audio device. The audio device may select the one of the voice recognition models for the expected trigger phrase based on a type of a companion device that is communicatively coupled to the audio device.

Abstract: A method for predicting the behavior of an electronic social network (ESN) includes identifying one user's connections with other users and creating a data structure in a memory that represents the users and their connections in the ESN. A plurality of data sources for electronic communications between users are analyzed and assigned a relative importance value. A weight is also assigned to each of the connections between the users. The weight is an encoded value computed based on a link structure of the connections where the link structure includes metadata indicating a category and a status of the respective connection. The probability that one user will communicate with one of the other users is calculated based on the analyzed plurality of data sources calculating, and the user's connections with respect to other users are ranked based on the calculated probabilities.

Abstract: One or more computer processors parse a received natural language question into an abstract meaning representation (AMR) graph. The one or more computer processors enrich the AMR graph into an extended AMR graph. The one or more computer processors transform the extended AMR graph into a query graph utilizing a path-based approach, wherein the query graph is a directed edge-labeled graph. The one or more computer processors generate one or more answers to the natural language question through one or more queries created utilizing the query graph.

Abstract: Methods, systems, and computer program products for explaining cross domain model predictions are provided herein. A computer-implemented method includes providing a test data point to a domain adaptation model to obtain a prediction, wherein the domain adaptation model is trained on a set of labeled data points and a set of unlabeled data points. The method includes generating a task specific explanation for the prediction that includes one or more data points from among the sets that satisfy a threshold score for influencing the prediction. Additionally, the method includes generating a domain invariant explanation for the prediction. The domain variation explanation is generated by ranking pairs of data points based on a statistical similarity to the test data point, wherein each pair includes a data point from the set of labeled data points and a data point from the set of unlabeled data points.