International Business Machines Corporation

Yishai Shimoni

Ehud Karavani

Peter Bak

Yoel Shoshan

Roie Melamed

Vadim Ratner

Matan Yitshak Ninio

Sivan Ravid

In some examples, a system for implementing a protocol for data transmission can include a processor to transmit a public key or an encrypted name to a ledger of a source computing device and receive a scoring code, training data, and data definitions from the source computing device. The processor can also transmit predicted labels encrypted with the public key to the source computing device, the predicted labels based on the training data and the data definitions, and receive labeled test data from the source computing device. The processor can also transmit decrypted predicted labels to the source computing device, receive a list of evaluation results from a set of remote computing devices, transmit scoring results based on the list of evaluation results to the set of remote computing devices, and execute an instruction based on a techniques corresponding to a highest score from the scoring results.

Distributed anonymous scoring technique

An example computer-implemented method includes receiving, via a processor, a dataset of objects to be tested for positivity violations. The method includes generating, via the processor, a feature matrix based on features extracted from the objects. The method also includes generating, via the processor, decision trees to group the objects into homogenous groups based on entropy and generate a random forest to compute consistency of violations. The method further includes detecting, via the processor, a positivity violation based on an entropy threshold, a consistency threshold, and a statistical significance threshold being exceeded. The method further includes generating, via the processor, an interactive representation to display the positivity violation.

Detecting positivity violations in multidimensional data

There is provided a method for unsupervised training of a machine learning model, comprising: receiving 3D images depicting a respective object, for each respective 3D image: dividing the 3D image into 3D patches, computing a first 2D image corresponding to a first orientation of the respective object, computing a second 2D image corresponding to a second orientation, automatically labelling pairs of 2D patches from the first and second 2D images with a patch measure indicative of likelihood of a certain 3D patch of the 3D image corresponding to a certain pair of 2D patches, training the ML model using a training dataset including the labelled patch pairs, for receiving patches extracted from first and second 2D images captured by an imaging sensor at the first and second orientations, and outputting an indication of likelihood of a visual finding in a 3D region of the object corresponding to the 2D patches.

Machine learning model for analysis of 2D images depicting a 3D object

Growing community of inventors

Givatayim, Israel

Yishai Shimoni