Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support cooperative training of machine learning (ML) models that preserves privacy in untrusted environments using distributed executable file packages. The executable file packages may include files, libraries, scripts, and the like that enable a cloud service provider configured to provide server-side ML model training to also support cooperative ML model training with multiple clients, particularly for a fraud prediction model for financial transactions. Because the cooperative training includes the clients training respective ML models and the server aggregating the trained ML models, private client data such as financial transaction data may be used to train the fraud prediction model without exposing the client data to others. Such cooperative ML model training enables offloading of computing resource-intensive training from client devices to the server and may train a more robust fraud detection model.

Abstract: The proposed systems and methods apply natural language processing to identify implicit security requirements flowing from input text narratively describing desired features for a software project. These systems and methods can identify hidden security requirements that may not be readily apparent from the features described in the input text. For example, a story may include a feature of a return URL (Uniform Resource Locator), which is the URL for the website to which a user will be redirected. A security vulnerability that would not be obvious from this feature is that a user might be directed to an attacker controlled site instead of the originally intended site. A security requirement that could counteract this vulnerability would be to include the feature of verifying all redirects go to Whitelisted Sites.

Abstract: A method and system for assisting program code development are disclosed. The method may include obtaining multimodal data of a meeting discussing a program code development, extracting a plurality of topics and a plurality of concepts from the multimodal data, identifying a plurality of meeting segments for the plurality of concepts. The method may further include determining a coding intent from program codes for the program code development, aligning the coding intent to a set of topics, identifying the set of concepts associated with the topic aligned with the coding intent, identifying a set of meeting segments associated with the concept. The method may further include determining an alignment metric of the meeting segment based on an alignment metric between the coding intent and the topic, and outputting one or more meeting segments for the coding intent based on alignment metrics of the one or more meeting segments.

Abstract: In some implementations, a scheduling platform may receive task information regarding a set of tasks for execution using a set of computing resources, wherein the task information includes, for the set of tasks, at least one of: a run time parameter, a priority parameter, or a success rate parameter. The scheduling platform may communicate with a computing resource management device to obtain first computing resource information regarding the set of computing resources. The scheduling platform may generate a first assignment of the set of tasks to the set of computing resources. The scheduling platform may transmit assignment information identifying the first assignment. The scheduling platform may receive second computing resource information. The scheduling platform may generate a second assignment of the set of tasks to the set of computing resources. The scheduling platform may transmit second assignment information identifying the second assignment.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer-storage media, for system contextualization. For example, a method can include obtaining information pertaining to (i) parts of a manufacturing plant and (ii) steps of a process executing on the manufacturing plant; generating one or more virtual representations associated with each of the parts and the steps of the process; receiving a query from a user, the query being in conformation with a format of a hierarchical template model; and generating a response to the query based on the one or more virtual representations of the manufacturing plant.

Abstract: Systems and methods for detecting, classifying, and managing impediments are disclosed. For example, embodiments may be related to impediments in project management. The proposed systems and methods are configured to evaluate data harvested from multiple different sources (in different formats), identify potential impediments that may be described or present in the data, and classify said impediments based on whether the impediment is non-technical or technical. In addition, the proposed systems implement a technical solution of active learning combined with reinforcement learning to produce a feedback loop that, over each iteration, improves the accuracy of the impediment classification. The impediment management assistant is configured to identify impediments from various inputs sources across industries with an AI-based self-learning capability, providing a robust and accurate model even with only a limited training dataset.

Abstract: Implementations include receiving graph data representative of a process-aware analytical attack graph (AAG) representing paths within an enterprise network with respect to observed facts of the enterprise network, the process-aware AAG at least partially defining a digital twin of the enterprise network, receiving data indicating at least one non-observed fact of the enterprise network, generating, from the graph data and the received data, an augmented process-aware AAG representing paths within the enterprise network with respect to the observed facts and the at least one non-observed fact, determining, by a process-aware risk assessment module, a risk assessment based on the augmented process-aware AAG, and providing, by a mitigation simulator module, a mitigation list based on the process-aware AAG and the risk assessment, the mitigation list comprising a prioritized list of observed facts of the process-aware AAG.

Abstract: The present disclosure relates to systems, methods, and products for using machine-learning networks to generate trustworthy audio and face mesh. A system, serving as a digital avatar, generates a trust audio and trust face mesh corresponding to an input text. A method includes generating a set of trust embedding vectors based on a reference audio; generate a text embedding vector based on the input text; generate a conditioned vector based on the set of trust embedding vectors and the text embedding vector; synthesize an audio representation based on the conditioned vector; generate the trust audio based on the synthesized audio representation; obtain a speech feature representation based on the trust audio; obtain an abstract feature vector based on the speech feature representation; and generate positions of vertices based on the abstract feature vector, the positions of vertices being used for generating the trust face mesh.

Abstract: In some implementations, an advisor system may receive a description of a problem to be solved and problem data identifying quantum computing-related and classical computing-related problems. The advisor system may perform natural language processing on the description of the problem and the problem data to respectively generate a problem embedding vector for the problem and to generate embedding vectors that represent the quantum computing-related and classical computing-related problems. The advisor system may process the problem embedding vector and the embedding vectors, with a vector matching model, to determine a semantically closest matching one of the embedding vectors to the problem embedding vector and, accordingly, may generate a recommendation that includes an indication to solve the problem with a classical computing resource, a quantum computing resource, or a combination of a classical computing resource and a quantum computing resource.

Abstract: In some implementations, a robot host may receive a video associated with assembly using a plurality of sub-objects. The robot host may determine spatio-temporal features based on the video and may identify a plurality of actions represented in the video based on the spatio-temporal features. The robot host may map the plurality of actions to the plurality of sub-objects to generate an assembly plan and may combine output from a point cloud model and output from a color embedding model to generate a plurality of sets of coordinates corresponding to the plurality of sub-objects. The robot host may perform object segmentation to estimate a plurality of grip points and a plurality of widths corresponding to the plurality of sub-objects. Accordingly, the robot host may generate instructions, for robotic machines, based on the assembly plan, the plurality of sets of coordinates, the plurality of grip points, and the plurality of widths.

Abstract: In some implementations, a user device may transmit, to a first data source, a first message encoding a first query and may receive a first response to the first query from the first data source. The user device may transmit, to a second data source, a second message encoding a second query that duplicates the first query and may receive a second response to the second query from the second data source. The user device may transmit a first reverse query to the first data source based on the second response and/or a second reverse query to the second data source based on the first response. The user device may receive at least one reverse query response. Accordingly, the user device may determine whether an answer to the first query is valid, based on the first response, the second response, and the at least one reverse query response.

Abstract: Implementations include providing a robotic model representative of a robotic system, determining baseline data by executing a simulation using the robotic model and a simulation world that represents the environment, generating, by a perturbation system, perturbations to be included in a set of test cases, each test case including at least one perturbation to at least one object relative to a baseline of the at least one object determined from the baseline data, determining test result data for the set of test cases by modifying the at least one object in view of the at least one perturbation to provide a modified object and executing the simulation with the modified object, and displaying a graph representative of the at least one object based on the baseline data and test result data for a test case, the graph depicting variance of the object during the simulation using the test case.

Abstract: Implementations include a computer-implemented method for reducing cyber-security risk, comprising: selecting one or more modules for inclusion in a knowledge mesh, wherein each module is associated with a respective aspect and maintains a knowledge graph specific to the respective aspect, wherein each knowledge graph is generated using data from one or more cyber-security repositories and includes nodes and connections between the nodes; receiving a query corresponding to a first node of a first knowledge graph included in the knowledge mesh; generating a response to the query by identifying connections between the first node of the first knowledge graph and at least one node of at least one other knowledge graph included in the knowledge mesh; and identifying, based on the response to the query, one or more actions to reduce cyber-security risk.

Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support functionality for identifying, quantifying, and mitigating risks. A dataset that includes a plurality of data items associated with one or more entities is generated and a set of rules is executed against the dataset to determine a sentiment for each data item of the plurality of data items. The dataset may be evaluated against a machine learning model configured to produce a set of risk categorizations that associate each of the plurality of data items with a risk category. Scoring metrics are generated for each of the data items based at least in part on the evaluating. The scoring metrics may account for an impact of dependencies and capabilities of the one or more entities on risks corresponding to the risk categorizations.

Abstract: Implementations include receiving demand data representing types of material and quantities of material demanded for each of a plurality of locations within a geographical area; receiving vessel data representing availabilities and capacities of each of a plurality of vessels; processing the demand data and the vessel data through a capacity optimization model to provide a first output comprising initial voyage plans for the plurality of vessels; receiving weather data representing predicted weather conditions within the geographical area; and processing the first output and the weather data through a sequence optimization model to provide a second output comprising updated voyage plans. Each initial voyage plan and each updated voyage plan defines a type of material, quantity of material, vessel, and sequence of locations.

Abstract: Implementations are directed to receiving, by an application, query data, the query data indicating a conditional query and a selection of files on which to execute the conditional query; identifying a subset of selected files for which copies of the selected files are not stored in a file database of the application; extracting, from a file system, a copy of each file of the subset of selected files into the file database of the application; parsing the copies of the selected files from the file database into a dataframe; executing the conditional query on the selected files in the dataframe; and providing a result of the conditional query. Actions include executing, by a discovery engine, a discovery process to provide metadata comprising, for a file in the file system: a name of the file; corresponding column names and data types; and a file alias.

Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support automated source code review using sentiment analysis with magnitude of entities. Known compliant and non-compliant source code may be used to generate dictionaries for evaluating lines of code using AI and ML techniques, such as by clustering data entities (lines of software code) and performing sentiment analysis on the data entities (lines of software code) which accounts for a magnitude of the data entities in the software code. The dictionaries enable automated review and correction of non-compliant code, such as vulnerable or insecure code, during the coding process. For example, sentiment analysis may be performed using the dictionaries on in-development code to determine a polarity and magnitude score for each line of code. The scores for each line can be compared to one or more conditions to determine a remediation action for individual lines of code.

Abstract: In some implementations, a font generator device may receive input text that includes one or more characters. The font generator device may generate a disability-oriented font based on one or more parameters that relate to a vision deficiency, wherein the disability-oriented font is readable by users that have the vision deficiency and unreadable by users that do not have the vision deficiency. The font generator device may transform the one or more characters included in the input text into the disability-oriented font. The font generator device may generate an output that represents the input text using the disability-oriented font based on transforming the one or more characters included in the input text into the disability-oriented font.

Abstract: Implementations include a computer-implemented method for reducing cyber-security risk, comprising: accessing a knowledge mesh including a plurality of modules, wherein each module is associated with a respective aspect and maintains a knowledge graph specific to the respective aspect, wherein each knowledge graph is generated using data from one or more cyber-security repositories and includes nodes and connections between the nodes; performing an information completion process to generate connections between nodes of knowledge graphs maintained by different modules of the knowledge mesh, including performing at least one of: inheritance-based inference; natural language processing classifier-based inference; or natural language processing-based object matching inference; and identifying, using the generated connections between the nodes of the knowledge graphs, one or more actions to reduce cyber-security risk.

Abstract: Aspects of the present disclosure provide systems, methods, apparatus, and computer-readable storage media that support improved watermarking and fingerprinting of a shared dataset. To illustrate, clustering may be performed on the dataset using initial clustering parameters (e.g., a secret key) to assign each record (e.g., attribute) of the dataset to one of multiple clusters. The secret key may be selected by a user or determined automatically based on the clustering algorithm. After the clustering, the records of each cluster may be selected for embedding a portion of fingerprint data based on one or more security parameters (e.g., a hash function, priority values, even/or selection, etc.). The selected records (or portions thereof) may be replaced with corresponding portions of the fingerprint data to embed the fingerprint data within different records as watermarking. Aspects also include analyzing a dataset to verify whether watermarking is present and to extract a fingerprint.