Google Inc.

Other

Steven Ross

Henry Will Schneiderman

Michael Nechyba

Yong Zhao

Christopher Farrar

Ronald A Hillock

Brian Patrick Williams

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.

Facial recognition can be used to determine whether or not a user has access privileges to a device such as a smartphone. It may be possible to spoof an authenticated user's image using a picture of the user. Implementations disclosed herein utilize time of flight and/or reflectivity measurements of an authenticated user to compare to such values obtained for an image of an object. If the time of flight distance and/or reflectivity distance for the object match a measured distance, then the person (i.e., object) may be granted access to the device.

Differentiating real faces from representations

An example method includes capturing, by an image capture device of a computing device, an image of a face of a user. The method further includes detecting, by the computing device, whether a distance between the computing device and an object represented by at least a portion of the image is less than a threshold distance, and, when the detected distance is less than a threshold distance, denying authentication to the user with respect to accessing one or more functionalities controlled by the computing device, where the authentication is denied independent of performing facial recognition based at least in part on the captured image.

Facial recognition

An apparatus and process for mar free handling of sheet steel pieces and blanks die cut therefrom in a press, in which load and unload trolleys are used with vacuum pickup and transport of the sheet steel piece and blanks from respective locations at the front and rear of the press. A pusher slides the sheet steel pieces off a low friction platform at the front of the press and onto a non scuffing but smooth and soft bed liner, and the cut blank is raised with the die by an array of vacuum cups each partially recessed into the die board. A table is extendible from the rear of the press to be positioned beneath the die and receive the blank thereafter released from the die cavity. The table is retracted to allow pick-up therefrom by the unload trolley.

Method and apparatus for mar free handling of sheet steel

A steel rule die is disclosed having strengthening layers bonded to the die board to stiffen the die board against flexing during handling, while alleviating the tendency to loosen the steel rule during handling and enhancing the durability of the die. KEVLAR graphite fiber fabric is impregnated and bonded to either face of the die board with epoxy resin.

Growing community of inventors

Allison Park, PA, United States of America

Steven Ross