Abstract: The proposed systems and methods provide a fixed set of intelligent, general APIs to manage access to enterprise data stored in a cloud-based data lake. These systems and methods allow a fixed set of APIs to respond to all queries regarding the stored enterprise data by using a cached reference table that locates the container and document in which the requested data is held. The proposed systems and methods provide a framework for a minimal API service code with the capacity for responding to dynamic queries while maintaining stringent privacy control protections.

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: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support optimization of communications transmitted over a plurality of communication mediums. Historical communications data may be analyzed to identify clusters of users and a model may be constructed based on the clusters. Candidate sequences of communications over a period of time (e.g., sequences of communication successfully triggering events) are identified using metrics (e.g., probabilities, attribution penalties, etc.) derived from the model or other information. The candidate sequences of communications may be determined at a group or cluster level and then tuned or optimized (e.g., using transition sequences, harmonization, entity priors, etc.) for individual users to produce optimized sequences of communications. The optimized sequences of communications may then be transmitted to individual users according to each user's optimized sequence of communications.

Abstract: A device may process the surveillance video data to segment vehicles, and may utilize a segmentation guided attention network model with the vehicles to determine traffic density count data. The device may process an image segmentation map, with a regression analysis model, to derive traffic signal timing. The device may process the surveillance video data, with a deep learning model, to identify objects, and may utilize a YOLO model, with the objects, to determine object types. The device may utilize a curriculum loss model with the objects to determine crowd count data, and may process the surveillance video data, with a video analytics model, to identify first events. The device may process the surveillance video data, with a classifier and deep network models, to identify second events, and may process the determined information, with a dynamic text-based explanation model, to generate a text-based explanation and/or a failure prediction.

Abstract: The present disclosure relates to a system, a method, and a product for intent discovery. The system includes a processor in communication with a memory storing instructions. When the processor executes the instructions, the instructions are configured to cause the processor to: obtain documents comprising a set of utterances, extract the set of utterances from the documents, generate a set of utterance embeddings based on the set of utterances, clusterize the set of utterance embeddings to obtain a plurality of clusters, obtain a cluster label for each cluster, encode each document based on a number of times each utterance cluster identifier (ID) appears to obtain an encoded document, perform latent Dirichlet allocation (LDA) on the encoded documents to obtain K topics, and each topic corresponding to a list of key clusters with cluster IDs, and for each topic, replace the cluster IDs with the cluster labels.

Abstract: A contract platform may receive, from data structures, historical contract data; read, from the historical contract data, numerical categorical data; process the numerical categorical data, with one or more artificial intelligence models, to convert the numerical categorical data into vector format; generate a taxonomy in vector format based on the converted numerical categorical data; read, from the historical contract data, nonnumerical noncategorical data; process the nonnumerical noncategorical data, with the models, to convert the nonnumerical noncategorical data into vector format; generate a knowledge graph in vector format based on the converted nonnumerical noncategorical data.

Abstract: Embodiments provide systems, method, and computer-readable storage media for performing robotic tasks in a distributed and coordinated manner. A fleet of robots may use a sequence of messages to appoint supervisors for a set of tasks, where the supervisor robots are responsible for ensuring that their supervised tasks are completed by other robots of the fleet. The supervisors may solicit requests from other robots to perform available tasks and select a robot to perform an available task. Once a robot is appointed, the supervisor and the worker may use messaging sequences to monitor the status of the task and participating robots (e.g., the supervisor and the worker). The monitoring may enable the supervisor to detect a failed worker and enable other robots to detect a failed supervisor.

Abstract: In some implementations, a system may determine, based on a qualification model, a prediction output of an analysis of user information. The system may determine, based on a generator model, a plurality of counterfactual explanations associated with the prediction output and the user information. The system may cluster, according to a clustering model, the plurality of counterfactual explanations into clusters of counterfactual explanations. The system may select, based on a classification model, a counterfactual explanation from a cluster of the clusters of counterfactual explanations. The system may provide a request for feedback associated with the counterfactual explanation. The system may receive feedback data associated with the request for feedback. The system may update a data structure associated with the clustering model based on the feedback data and the counterfactual explanation to form an updated data structure. The system may perform an action associated with the updated data structure.

Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support detection and mitigation of defects occurring in outputs of a process. A modelling engine creates and trains models to identify defects occurring in outputs of a process or system. A set of causation rules may be created and trained to identify causes of different defects identifiable via the models and a set of control rules may be created and trained to generate control data that may mitigate the causes of identified defects. A process monitoring device incorporating the trained models and rules may then be used to monitor the process and detect defects using the models. Once a defect is detected, the rules may be applied to determine the cause of the defect and then control data may be generated to modify the process such that further occurrences of the defect are reduced or eliminated.

Abstract: A device may receive purchase data identifying purchases by users of user devices and identifying non-temporal data associated with the users, and may preprocess the purchase data to generate sequences of multivariate and multimodal symbols. The device may process the sequences of multivariate and multimodal symbols, with a long short-term memory based encoder-decoder model, to generate sequence embeddings, and may process the non-temporal data associated with the users, with a knowledge graph, to determine knowledge graph embeddings capturing the non-temporal data. The device may process the sequence embeddings and the knowledge graph embeddings, with a knowledge graph embedding model, to generate modified sequence embeddings, and may process the modified sequence embeddings, with a clustering model, to determine clusters of the users in relation to products or services purchased by the users. The device may perform one or more actions based on the clusters of the users.

Abstract: Methods, systems, and computer-readable storage media for receiving, by an identity network and from a verifying entity, a query including an identifier that uniquely identifies an entity that is active in the mobility network and an attestation that is to be authenticated, resolving, by the identity network, the query to provide a resolved query, the resolved query including the attestation, resolving at least partly including identifying a data source of a plurality of data sources that is to be queried to authenticate the attestation, transmitting, by the identity network, the resolved query to the data source, receiving, by the identity network, a response to the resolved query, and transmitting, by the identity network, the response to the verifying entity.

Abstract: Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support secure 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 ML model training based on non-encrypted data to also support homomorphic encryption of data and ML model training with one or more clients, particularly for a diagnosis prediction model trained using medical data. Because the training is based on encrypted client data, private client data such as patient medical data may be used to train the diagnosis prediction model without exposing the client data to the cloud service provider or others. Using homomorphic encryption enables training of the diagnosis prediction model using encrypted data without requiring decryption prior to training.

Abstract: Systems and methods for responsible AI compliance and governance management in AI Products are disclosed. The system receives a request to assess an enterprise product associated with a specific application. Further, the system may determine a plurality of datasets associated with the AI model of the enterprise product. Furthermore, the system generates a training dataset and a test dataset for the determined plurality of datasets associated with the AI model. The system generates a ranked list of recommended metrics for the enterprise product based on the generated training dataset and the test dataset. The system further determines a mitigation strategy for the enterprise product based on the generated ranked list of recommended metrics. Furthermore, the system creates a feedback loop for continuous training and tuning the AI model and the plurality of datasets based on the determined mitigation strategy.

Abstract: Method, data processing system, and computer-readable storage media for responding to a user query. Receiving query from user, query pertaining to request for information. Based on query, generate prompts by masking sensitive information in query. Receive responses from foundation models in response to inputting prompts. Based on responses, generate common result set. By validating common result set with sensitive information, generate response. By supplementing response with sensitive information, generate user response. Providing user response in response to query to the user.

Abstract: Methods, systems, and apparatus are provided for generating an image. A personalized text prompt is generated by processing an input embedding using a transformer model followed by a first fully connected neural network. The input embedding comprises a multi-dimensional embedding vector associated with a user profile and a plurality of user items. A scored label set is generated identifying a user's preferences by processing a set of attributes for the plurality of user items using a second fully connected neural network. The image is generated by processing the personalized text prompt and the scored label set using a diffusion model.

Abstract: Packaging devices such as bags, pouches, envelops, containers, and the like, can include an integrated heating system. The packaging devices described herein have multiple potential uses such as for the containment and heating of single-serving meals, drinks, massage oils, masks, body wax, anti-wrinkle eye creams, and the like, to be heated on the go. The packaging devices described herein can also be used for instant heating of food and/or beverages for backcountry use. In some embodiments, the packaging devices use fibrous natural materials (e.g., leaf skeletons and soft biomaterials like chitosan) along with silver nanowires to create completely biodegradable, and reusable self-heating packaging.

Abstract: Methods, systems, and apparatus for a bi-directional quantum annealing approach to Markov random field networks for machine learning in image analysis. In one aspect, a method includes obtaining training data comprising features extracted from a first set of images; training a deep quantum restricted Boltzmann machine (QRBM) comprising multiple layers using the training data, the training comprising layer-wise training of the multiple layers, wherein training each layer of the multiple layers comprises evaluating a restricted Boltzmann machine (RBM) probability distribution using bi-directional quantum annealing; and validating the trained deep QRBM using test data comprising features extracted from a second set of images.

Abstract: A system for issue prediction based on multidimensional data analysis includes a model generator that receives a resolved data item relating to a service issue. The resolved data item includes different attributes corresponding to multiple data dimensions and adjusts a population of attributes based on a statistical data model and a deep learning data model operating independent of each other. The statistical data model operates on the attributes for providing a predictive feature and the deep learning data model operates on the attributes for providing a predictive label based on performance metrics related to the data dimensions. The predictive feature and the predictive label collectively define training data. The model generator also trains a classification model based on the training data for predicting a potential issue related to an unresolved data item. The trained data model provides a trigger based on the potential issue being related to the performance metrics.

Abstract: A device may receive tracking data identifying trajectories of objects, and may annotate the tracking data to identify object categories for the tracking data. The device may identify objects based on the object categories, and may determine object trajectories for the objects. The device may transform the object trajectories to overhead planar trajectories, and may generate simplified object trajectories. The device may convert a continuous space associated with the simplified object trajectories into a discrete space, and may convert the discrete space to a lower dimensional space. The device may transform the object trajectories to binary feature vectors, and may process the binary feature vectors, with a clustering model, to determine trajectory clusters. The device may train a classification model with the binary feature vectors and the trajectory clusters, and may cause the trained classification model to be implemented.

Abstract: An Artificial Intelligence (AI) based filter apparatus includes an input filter and an output filter protecting a generative AI model and preventing restricted content from being transmitted to user devices. When a user query is received, the input filter determines if the user query can be transmitted to the generative AI model by generating an input risk score for the received user query. If the user query is transmitted and a model query response is received from the generative AI model, the output filter determines an output risk score based on which the model query response may be transmitted to the user. The input filter and the output filter each include a pre-trained language model as a base with additional layers trained to estimate the corresponding risk scores.