Oracle International Corporation

Daniel Peterson

Virendra J Marathe

Pallika Haridas Kanani

Jean-Baptiste Frederic George Tristan

Robert James Oberbreckling

Anshuman Suri

Subject level privacy attack analysis for federated learning may be performed. A request that selects an analysis of one or more inference attacks may be received to determine a presence of data of a subject in a training set of a federated machine learning model. The selected inference attacks may be performed to determine the presence of the data of subject in the training set of the federated machine learning model. Respective success measurements may be generated for the selected inference attacks based on the performance of the selected inference attacks, which may then be provided.

Subject level privacy attack analysis for federated learning

Herein are techniques for domain adaptation of a machine learning (ML) model. These techniques impose differential privacy onto federated learning by the ML model. In an embodiment, each of many client devices receive, from a server, coefficients of a general ML model. For respective new data point(s), each client device operates as follows. Based on the new data point(s), a respective private ML model is trained. Based on the new data point(s), respective gradients are calculated for the coefficients of the general ML model. Random noise is added to the gradients to generate respective noisy gradients. A combined inference may be generated based on: the private ML model, the general ML model, and one of the new data point(s). The noisy gradients are sent to the server. The server adjusts the general ML model based on the noisy gradients from the client devices. This client/server process may be repeated indefinitely.

Privacy preserving collaborative learning with domain adaptation

Systems and methods are disclosed to improve a topic modeling system that tunes a topic model for a set of topics from a corpus of documents, by allowing users to pre-inform the tuning process with bias parameters for desired associations in the topic model. In embodiments, the topic model may be a Latent Dirichlet Allocation (LDA) model. In embodiments, the bias parameter may indicate a fixed association where a particular word in a particular document is associated with a particular topic. In embodiments, the bias parameter may specify a weight value that biases the inference process with regard to a particular association. Advantageously, the disclosed features allow users to specify a small number of parameters to steer the tuning process towards a set of desired topics. As a result, the topic model may be generated more quickly and with more useful topics.

Growing community of inventors

Broomfield, CO, United States of America

Daniel Peterson