International Business Machines Corporation

Skyler Speakman

Komminist S Weldemariam

Celia Cintas

Girmaw Abebe Tadesse

David W Kaguma

Isaac Waweru Wambugu

Eric Mibuari

Felix Bonard Kwizera

Juliet K Mutahi

Violette Ogega

Kush Raj Varshney

Supriyo Chakraborty

Karthikeyan Natesan Ramamurthy

Victor Abayomi Akinwande

Srihari Sridharan

Dennis Wei

William Ogallo

Julian Bertram Kuehnert

Amanda Coston

Zairah Mustahsan

Described are techniques for A/B testing including a computer-implemented method of identifying, in an A/B testing database, a set of feature values with a statistically significant difference in A/B testing outcomes above a threshold. The method further includes partitioning records in the A/B testing database into a plurality of population strata according to the set of feature values. The method further includes performing A/B testing, and identifying heterogeneous outcomes of the A/B testing for respective strata of the plurality of population strata.

Subgroup analysis in A/B testing

An illustrative embodiment includes a method for detecting whether a given item within input data is anomalous. The method includes: extracting activations at respective nodes of an autoencoder trained on the data, the activations comprising activations at the respective nodes for the given item within the data and for one or more other items within the data; calculating p-values corresponding to the respective nodes, wherein the p-value for a given node represents a proportion of the activations at the given node for the other items which are greater than the activations at the given node for the given item; determining at least one score at least in part by scanning for a subset of the respective nodes which maximizes a statistical scoring function applied to the corresponding p-values; and providing a visualization of at least the given item within the input data to a human user.

Identifying adversarial attacks with advanced subset scanning

Embodiments of the present invention provide a computer system, a computer program product, and a method that comprises generating a baseline prediction for each database in a plurality of databases, wherein the generated baseline prediction includes datapoints for each performance feature of a respective database in the plurality of databases; generating a linkage matrix using a dimension reduction algorithm in an equal-dimensional feature space based on a measured difference between actual performance datapoints of each respective database and respective datapoints associated with generated baseline prediction for each database in the plurality of databases; linking one or more databases in the plurality of databases based on the measured difference within the generated linkage matrix reaching a predetermined threshold of measured performance; and generating a design using the linked databases for a computer system capable of being displayed through a user interface.

Automated data linkages across datasets

A method of utilizing a computing device to correct source data used in machine learning includes receiving, by the computing device, first data. The computing device corrects the source data via an application of a covariate shift to the source data based upon the first data where the covariate shift re-weighs the source data.

Enhancing fairness in transfer learning for machine learning models with missing protected attributes in source or target domains

The disclosure provides systems and methods for analysis of financial data and customization of financial products. Time-dependent data is obtained for a relatively large set of users, and from the data are determined a relatively small set of trajectories that model user behaviour. New users are characterized according to the trajectories and financial products are automatically personalised to suit the new user.

Methods and systems for offering financial products

A request to install an application on a device may be received, and data associated with the device and a set of users associated with the device may be received. Acceptance factors specified in a terms and conditions document associated with the application to be installed on the device may be identified. A terms and conditions implication of installing the application on the device may be determined based on the acceptance factors. Based on the terms and conditions implication, a rule may be dynamically generated to control at least a running of the application on the device. The rule may be activated or caused to be activated on the device. The activation of the rule may control the running of the application on the device.

Controlling computer executions based on terms and conditions and user device attributes

Herein are methods and systems for automated assessment of user devices and associated data, and granting access to wireless networks based on the assessment. For example, the systems and methods assess a user request for access to a wireless network at an access point, and automatically determine the type of access to be granted based on at least a loyalty score associated with the user. The automatic assessment can be used to encourage customer loyalty as well as to manage limited network resources.

Automated granting of access to networks and services for pre-paid devices

Growing community of inventors

Nairobi, Kenya

Skyler Speakman