Paypal, Inc.

Yaqin Yang

Fransisco Kurniadi

Dinesh Kumar

Yang Wu

Nitin S Sharma

Jiadi Xiong

Systems and methods for integrated multi-factor multi-label analysis include using one or more deep learning systems, such as neural networks, to analyze how well one or more entities are likely to benefit from a targeted action. Data associated with each of the entities is analyzed to determine a score for each of the proposed targeted actions using multiple analysis factors. The scores for each analysis factor are determined using a different multi-layer analysis network for each analysis factor. The scores for each analysis factor are then combined to determine an overall score for each of the proposed targeted actions. The entities and the proposed targeted actions with the highest scores are then identified and then used to determine which entities are to be the subject of which targeted actions.

Systems and methods for integrated multi-factor multi-label analysis

A system, a medium, and a method are provided to exchange data packets over a communications network and perform machine learning operations. A network server device receives account data from client devices that correspond to account profiles. An account engine of the network server device segments the account profiles into profile groups based on a respective balance associated with each account profile. The account engine determines target accounts from profile groups based on behavioral data. Further, data processing components of the network server device determine a method of contact for each target account. The data processing components determine a respective time to communicate with a respective device for each target account. Further, communication components of the network server device initiate communications to the respective devices at the respective times for each target account.

Server device configurations based on machine learning

Software architectures relating to machine learning (e.g., relating to classifying sequential text data. Unlabeled sequential text data may be produced by a variety of sources such as text messages, email messages, message chats, social media applications, and web pages. Classifying such data may be difficult due to the freeform and unlabeled nature of text data from these sources. Thus, techniques for training a machine learning algorithm to classify unlabeled text data in freeform format. Training is based on generation of labelling functions from lexical databases, applying the labelling functions to unlabeled text data in an unsupervised manner, and generating trained classifiers that accurately classify the unlabeled text data. The trained classifiers may then be implemented classify text data accessed from the variety of sources. The present techniques provide high-quality and efficient labeling of unlabeled text data in freeform formats.

Classifier determination through label function creation and unsupervised learning

Systems and methods for data management using machine learning and artificial intelligence techniques related to topic modeling on text comments are described. The text comments may correspond to a particular transaction conducted by a user. Machine learning text analysis is performed on the text comment to determine one or more topics associated with the text comment. The topic with the highest correlation to the text comment is assigned to the transaction claim. Based on the topic assigned to the transaction claim, various actions may be performed, including remedial actions on a user account. These techniques may be applicable to chargeback fraud, in some embodiments.

Data management using topic modeling

A system performs operations that include identifying a first subset of accounts from a set of accounts, each account in the first subset of accounts satisfying a first abuse score threshold for a first time period, the first abuse score threshold corresponding to a first buyer abuse component. The operations further include determining a first restriction rate for the first subset of accounts based on a number of accounts in the first subset of accounts that have been restricted for potential abuse. The operations also include comparing the first restriction rate with respective restriction rates of one or more other subsets of the set of accounts that correspond to one or more other abuse components, and based on the comparing, determining whether to adjust the first abuse score threshold.

Growing community of inventors

San Jose, CA, United States of America

Yaqin Yang