Google Inc.

Christopher Farrar

Steven Ross

Implementations relate to providing, in response to a query, machine learning model output that is based on output from a trained machine learning model. The machine learning model output can include a predicted answer to the query, that is predicted based on the trained machine learning model. The machine learning model output can additionally or alternatively include an interactive interface for the trained machine learning model. Some implementations relate to generating a trained machine learning model 'on the fly' based on a search query. Some implementations additionally or alternatively relate to storing, in a search index, an association of a machine learning model with a plurality of content items from resource(s) on which the machine learning model was trained.

Generating and/or utilizing a machine learning model in response to a search request

A method for rejecting biased data includes receiving a bias training data set based on a probability distribution of bias-sensitive variables of a target population and segmenting the bias training data set into clusters based on at least one respective bias-sensitive variable of the target population, each cluster including a bias cluster weight. The method also includes receiving a training data set for a machine learning model and segmenting the training data set into training clusters. Each training cluster is associated with at least one corresponding bias-sensitive variable of the target population and includes a corresponding training data set weight. The method also includes adjusting each training data set weight to match a respective bias cluster weight to form an adjusted training data set and providing the adjusted training data set to the machine learning model as an unbiased training data set.

Rejecting biased data using a machine learning model

A method for rejecting biased data using a machine learning model includes receiving a cluster training data set including a known unbiased population of data and training a clustering model to segment the received cluster training data set into clusters based on data characteristics of the known unbiased population of data. Each cluster of the cluster training data set includes a cluster weight. The method also includes receiving a training data set for a machine learning model and generating training data set weights corresponding to the training data set for the machine learning model based on the clustering model. The method also includes adjusting each training data set weight of the training data set weights to match a respective cluster weight and providing the adjusted training data set to the machine learning model as an unbiased training data set.

Growing community of inventors

Wexford, PA, United States of America

Christopher Farrar