Accenture Global Solutions Limited

Anwitha Paruchuri

Guanglei Xiong

Sukryool Kang

Lan Guan

Jayashree Subrahmonia

Anupam Anurag Tripathi

Sujeong Cha

Yuan He

Neeraj D Vadhan

Thomas Wayne Hancock

Jill Gengelbach-wylie

Colin Connors

Neeru Narang

Christopher Yen-Chu Chan

Ditty Mathew

Tsunghan Wu

Louise Noreen Barrere

Surya Raghavendra Vadlamani

Wenxian Zhang

Andrew Francis Hickl

Emmanuel Munguia Tapia

Chinnappa Guggilla

Anshuma Chandak

Danielle Moffat

Ivan A Wong

Peter Royer Smith, Jr

Tapia Emmanuel Munguia

Armando Ferreira Gomes

Angela Yang Sanford

Jing Su Brewer

Aaron James Sander

Ivan Wong

The present disclosure relates to a system, a method, and a product for using machine learning models to quantify and/or improve trust in conversations. The system includes a non-transitory memory; and a processor in communication with the non-transitory memory. The processor executes the instructions to cause the system to: obtain a set of vocal features and a set of text features for each sample in audio samples; obtain a trust score for each sample; perform a preprocess to obtain a set of input features for each sample; determine a type of machine-learning algorithm for the machine-learning network; tune a set of hyper parameters for the machine-learning network; generate a predicated trust score by the machine-learning network with the sets of input features for each sample; and train the machine-learning network based on the predicated trust score and the trust score for each sample to obtain the training result.

Systems and methods to improve trust in conversations

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.

Systems and methods to automate trust delivery

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.

Hyper-personalized prompt based content generation

Implementations are directed to receiving a set of tuples, each tuple including an entity and a product from a set of products, for each tuple: generating, by an embedding module, a total latent vector as input to a recommender network, the total latent vector generated based on a structural vector, a textual vector, and a categorical vector, each generated based on a product profile of a respective product and an entity profile of the entity, generating, by a context integration module, a latent context vector based on a context vector representative of a context of the entity, and inputting the total latent vector and the latent context vector to the recommender network, the recommender network being trained by few-shot learning using a multi-task loss function, and generating, by the recommender network, a prediction including a set of recommendations specific to the entity.

Few-shot learning for multi-task recommendation systems

The present disclosure relates to a system, a method, and a product for using deep learning models to quantify and/or improve trust in conversations. The system includes a non-transitory memory storing instructions executable to construct a deep-learning network to quantify trust scores; and a processor in communication with the non-transitory memory. The processor executes the instructions to cause the system to: obtain a trust score for each voice sample in a plurality of audio samples, generate a predicated trust score by the deep-learning network based on each voice sample in the plurality of audio samples, wherein the deep-learning network comprises a plurality of branches and an aggregation network configured to aggregate results from the plurality of branches, and train the deep-learning network based on the predicated trust score and the trust score for each voice sample to obtain a training result.

Systems and methods to improve trust in conversations with deep learning models

A system and method for lead conversion using conversational virtual avatar is disclosed. System comprising processor causes Conversation Virtual Avatar Platform (CVAP) to receive, for first entity, from lead prioritization engine, leads applicable to first entity via lead repository based on scores associated with respective leads. Processor causes CVAP to receive, through conversation management engine (CME) configured in CVAP, from leads, responses to questions pertaining to product attributes and information pertaining to lead. The processor causes CVAP to process responses to determine action and/or state, which includes whether to issue additional product-attribute based questions through Virtual Avatar (VA) using Response To Motion Module (RTME) or to, through recommender engine that uses recommendation model, in real-time, recommend products associated with first entity to respective at least one lead based on any or combination of responses received from lead, information pertaining to lead, and products ordered by entities similar to lead.

System, method, and medium for lead conversion using a conversational virtual avatar

A document processing system processes documents including typewritten and/or handwritten data by converting them to document images for entity extraction. A received document is initially processed to generate a deep document data structured and for classification as one of a structured or an unstructured document. If the document is classified as a structured document, it is processed for entity extraction based on a matching template and image alignment of the document image with the matching template. If the document is classified as an unstructured document, entities are extracted by obtaining nodes and providing the nodes to a self-supervised masked visual language model.

Deep document processing with self-supervised learning

Implementations of the present disclosure include receiving a goal, providing a problem-specific knowledge graph that is responsive to at least a portion of the goal, determining a set of events from the problem-specific knowledge graph, processing data representative of events in the set of events through a first machine learning (ML) model to provide a set of event scores, each event score in the set of event scores being associated with a respective event in the set of events, determining a sub-set of events based on the set of event scores, for each event in the sub-set of events, determining at least one action by processing a sequence of actions through a second ML model, and outputting the sub-set of events and a set of actions for execution of at least one action in the set of actions.

Early pattern detection in data for improved enterprise operations

The proposed systems and methods describe an autonomous asset detection system that leverages artificial intelligence (AI) models for three-dimensional asset identification and damage detection, asset damage classification, automatic in-field asset tag readings, and real-time asset management. In some embodiments, a deep learning-based system receives a set of aerial images of one or more assets and automatically identifies each asset in the image(s) using rotational coordinates. In some embodiments, an image annotation tool labels the images either manually or automatically. The system then detects whether the asset is damaged and, if so, determine the type of damage, and further captures and stores asset tag information for the target asset. The collected and processed data is then provided to end-users via a comprehensive user interface platform for managing the assets in real-time. The platform can also be configured to generate insights on damage and repair, facilitating an expedited and thoughtful response to asset maintenance.

System and method for identifying structural asset features and damage

A attribute-based data matching system determines top matches for a first element from a plurality of second elements. The data matching system extracts attributes for the first element from first dataset and attributes for the plurality of second elements from a plurality of second datasets. Pairs of attributes are generated wherein each attribute pair includes an attribute of the first element and an attribute of one of the plurality of second elements. Pairwise rankings of the plurality of second elements corresponding to the attributes of the first element are generated based on weights of the attribute pairs. The pairwise rankings of the attribute pairs are aggregated to determine a ranked list that orders the plurality of second elements based on the extent of their match with the first element. User feedback to the ranked list can be collected and used to adjust the data matching system.

Modular attribute-based multi-modal matching of data

A data extraction and expansion system receives documents with data to be processed, extracts a set of a specific type of entities from the received documents, expands the set of entities by retrieving additional entities of the specific type from an ontology and other external data sources to improve the match between the received documents. The ontology includes data regarding entities and relationships between entities. The ontology is built by extracting the entity and relationship information from external data sources and can be constantly updated. If the additional entities to expand the set of entities cannot be retrieved from the ontology then a real-time search of the external data sources is executed to retrieve the additional entities from the external data sources.

Section segmentation based information retrieval with entity expansion

Growing community of inventors

San Jose, CA, United States of America

Anwitha Paruchuri