Oracle International Corporation

At&t Intellectual Property I, LP

Sudhakar Kalluri

Venkata Chandrashekar Duvvuri

Maisam Shahid Wasti

Swetha Krishnakumar

Samba Reyes Njie

The present disclosure generally relates to evaluating communication workflows comprised of tasks using machine-learning techniques. More particularly, the present disclosure relates to systems and methods for generating a prediction of a task outcome of a communication workflow, generating a recommendation of one or more tasks to add to a partial communication workflow to complete the communication workflow, and generating a vector representation of a communication workflow.

Enhanced processing for communication workflows using machine-learning techniques

Techniques for summarizing lists for machine learning operations are disclosed. In some embodiments, a machine learning system generates feature vectors for a set of items based on varying values among a set of feature attributes. The system further generates, based on the feature vectors a set of clusters and generates a summary vector for a list of items as a function of the distribution of the items within the set of clusters, where the summary vector has a length equal to how many clusters are in the set of clusters. Summary vectors may be generated for a plurality of examples within a training dataset. The system may use the summary vectors to train a machine learning model to estimate unknown labels for new examples.

Systems and methods for optimizing machine learning models by summarizing list characteristics based on multi-dimensional feature vectors

Techniques are described herein for training and evaluating machine learning (ML) models for document processing computing applications using generalized vocabulary tokens. In some embodiments, an ML system determines a set of tokens for non-textual content in a plurality of documents. The ML system generates a fixed-length vocabulary that includes the set of tokens for the non-textual content. The ML system further generates for each respective document in a training dataset of documents, a respective feature vector based at least in part on which tokens in the fixed-length vocabulary occur in the respective document. The ML system trains a ML model based at least in part on the respective feature vector for each respective document in the training dataset.

Systems and methods for training and evaluating machine learning models using generalized vocabulary tokens for document processing

Techniques are described for training machine learning (ML) models using one or more electronic lists of items previously used in campaigns and labeled with an engagement rate corresponding to the list. A vocabulary formed from a union of the one or more lists may then be used to generate at least some items of a target recipient list. An engagement rate for the target recipient list may be inferred for the target recipient list. Natural language processing (NLP) techniques may be also be applied to optimize an engagement rate of a target recipient list and/or select content for the list.

Systems and methods for analyzing a list of items using machine learning models

Techniques are described herein for training and evaluating machine learning (ML) models for document processing computing applications based on in-domain and out-of-domain characteristics. In some embodiments, an ML system is configured to form feature vectors by mapping unknown tokens to known tokens within a domain based, at least in part, on out-of-domain characteristics. In other embodiments, the ML system is configured to map the unknown tokens to an aggregate vector representation based on the out-of-domain characteristics. The ML system may use the feature vectors to train ML models and/or estimate unknown labels for the new documents.

Hybrid in-domain and out-of-domain document processing for non-vocabulary tokens of electronic documents

Disclosed are techniques for determining the impact of including a token-set (e.g., text in the form of unigrams, bigrams, or trigrams) in a communication on a target outcome. More particularly, the present disclosure relates to techniques for determining the impact of the token-set based on, for example, the token-sets included in previous communications transmitted to user devices and the corresponding user responses to those previous communications.

Techniques for selecting content to include in user communications

Aspects of the subject disclosure may include, for example, applying a topic detection process to documents to obtain automatically detected topics and groups of automatically detected words, comparing the automatically detected topics with manually determined topics to determine actual purity metrics, determining an error metric based on a measure of deviation between ideal purity metrics and the actual purity metrics, and adjusting a parameter of the topic detection process according to the error metric resulting in an adjusted topic detection process. Other embodiments are disclosed.

Method and device for adjusting and implementing topic detection processes

The present disclosure generally relates to techniques for determining a segment of a content distribution plan. More specifically, the present disclosure discloses techniques for determining one or more key-value pairs of a content distribution plan by leveraging a trained machine learning model. A plurality of electronic communications may be generated based on completed key-value pairs with a content distribution plan. The plurality of electronic communications may then be distributed to a plurality of devices within a networked environment.

Growing community of inventors

Cupertino, CA, United States of America

Sudhakar Kalluri