Paypal, Inc.

Other

Nitin S Sharma

Michael Charles Todasco

Abhishek Chhibber

Darshankumar Bhadrasinh Desai

Vidyut Mukund Naware

Fransisco Kurniadi

Yaqin Yang

Yang Wu

Hubert Andre Le Van Gong

Grahame Andrew Jastrebski

Hong Fan

Vishal Sood

Sudhindra Murthy

Lov Kher

Ashwin Maruti Hegde

Mozhdeh Rouhsedaghat

Yaron Raps

Niraj Kumar

Quantum computers with a limited number of input qubits are used to perform machine learning processes having a far greater number of trainable features. A list of features of a field are divided into a plurality of feature groups. Each of the feature groups includes a respective group of some, but not all, of the features. A first machine learning process is performed to train a first instance of a quantum computer model, where the feature groups are used as inputs. Based on the first machine learning process being performed, a subset of the feature groups is selected for a second machine learning process. Thereafter, the second machine learning process is performed to train one or more second instances of the quantum computer model. The individual features of the selected subset of the feature groups are used as inputs for the second instances of the quantum computer model.

Quantum-computer-based machine learning

Techniques are disclosed in which a computing device repeatedly trains, using a stream of user data received at the computing device, a baseline model to generate a device-trained model, wherein the baseline model is trained at the computing device without providing user data included in the stream to a server computer system. In some embodiments, the computing device inputs, to the device-trained model, a set of characteristics associated with a user request received from a user of the computing device, wherein the device-trained model outputs a score for the user request. In some embodiments, the computing device transmits, to the server computer system, the score for the user request, wherein the transmitting includes requesting a decision for the user request. In some embodiments, the computing device performs an action associated with the user request in response to receiving a decision for the user request from the server computer system.

Edge device machine learning

Techniques are disclosed in which a computer system receives, from a plurality of user computing devices, a plurality of device-trained models and obfuscated sets of user data stored at the plurality of user computing devices, where the device-trained models are trained at respective ones of the plurality of user computing devices using respective sets of user data prior to obfuscation. In some embodiments, the server computer system determines similarity scores for the plurality of device-trained models, wherein the similarity scores are determined based on a performance of the device-trained models. In some embodiments, the server computer system identifies, based on the similarity scores, at least one of the plurality of device-trained models as a low-performance model. In some embodiments, the server computer system transmits, to the user computing device corresponding to the low-performance model, an updated model.

Federated machine learning management

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

Techniques are disclosed in which an edge user computing device pre-processes a stream of user data prior to using the stream of data to train a machine learning model at the edge device. The edge device receives the stream of user data, where the stream of data includes a first set of characteristics associated with the edge device and a second set of characteristics associated with a plurality of user requests received from a user of the edge device. The edge device repeatedly generates, using the stream of data, sets of pre-processed user data by performing pre-processing techniques on characteristics included in the stream of data. The edge device repeatedly trains, using the sets of pre-processed data, a baseline model to generate a device-trained model, where the baseline model is trained at the edge device without providing user data included in the stream of data to a server computer system.

Edge device representation learning

Techniques are disclosed for automatically retraining a machine learning model based on the performance of this model falling below a performance threshold. In some embodiments, a computer system compares output of a new machine learning model for a new set of examples with known labels for examples in the new set of examples, wherein the new set of examples includes one or more new features. In some embodiments, the computer system determines, based on the comparing, whether a current performance of the new machine learning model satisfies a performance threshold for machine learning models, where the performance threshold is based on output of a benchmark machine learning model. In some embodiments, the computer system automatically triggers, in response to determining that the current performance of the new model does not satisfy the performance threshold, retraining of the new model.

Meta-learning and auto-labeling for machine learning

Techniques are disclosed in which a computer system receives a transaction request and uses a federated machine learning model to analyze the transaction request. A server computer system may generate a federated machine learning model and distribute portions of the federated machine learning models to other components of the computer system including a user device and/or edge servers. In various embodiments, various components of the computer system apply transaction request evaluation factors to the portions of the federated machine learning model to generate scores. The server computer system uses the scores to determine a response to the transaction request.

Device-side federated machine learning computer system architecture

Systems, methods, and computer program products are directed to machine learning techniques that use a separate embedding layer. This can allow for continuous monitoring of a processing system based on events that are continuously generated. Various events may have corresponding feature data associated with at least one action relating to a processing system. Embedding vectors that correspond to the features are retrieved from an embedding layer that is hosted on a separate physical device or a separate computer system from a computer that hosts the machine learning system. The embedding vectors are processed though the machine learning model, which may then make a determination (e.g. whether or not a particular user action should be allowed). Generic embedding vectors additionally enable the use of a single remote embedding layer for multiple different machine learning models, such as event driven data models.

Optimally compressed feature representation deployment for automated refresh in event driven learning paradigms

Techniques are disclosed relating to training a machine learning model to handle adversarial attacks. In some embodiments, a computer system perturbs, using a set of adversarial attack methods, a set of training examples used to train a machine learning model. In some embodiments, the computer system identifies, from among the perturbed set of training examples, a set of sparse perturbed training examples that are usable to train machine learning models to identify adversarial attacks, where the set of sparse perturbed training examples includes examples whose perturbations are below a perturbation threshold and whose classifications satisfy a classification difference threshold. In some embodiments, the computer system retrains, using the set of sparse perturbed training examples, the machine learning model. The disclosed techniques may advantageously enable a machine learning model to correctly classify data associated with adversarial attacks.

Automatic machine learning vulnerability identification and retraining

A method for optimizing delivery of digital information packets across a network of linked, packet switching nodes is disclosed. A suitably programmed computer translates business oriented requests into network technical requirements. The computer then uses knowledge including pre-stored or real-time discovered topographical maps of the network, and a database of the current attributes of the components of the network, to automatically determine settings for software addressable forwarding devices on the network to implement the network technical requirements, and then configures the network remotely via the software addressable switches at nodes on the network.

Growing community of inventors

San Jose, CA, United States of America

Nitin S Sharma