Ganit Prisant Henkin
Intel (Israel)
Harvesting data and harnessing science for intelligent testing failures classification
Overview regarding AI solutions and data harvesting process in the testing context
Problem statement – initial triage of failing tests can be challenging and has high potential for time and efforts consuming as well as may cause failed test to disrupt or spend the time of the wrong people.
Proposed solution – Based on historical data, use machine learning for initial classifying of the different regression failures. The classification will be presented as a “recommendation system”, supplying and assisting the test regression owner a “Failure Reason/Area” with a calculated confidence/probability level
Description of a Proof Of Concept implemented in our team, their results, next steps and challenges
Generalization – to which other purposes can we use this type of data-based technique?
I have been a software engineer in embedded Real-Time environment for over 25 years. I have started my career at Freescale (former Motorola) Semiconductors, and I am currently a firmware micro-architect and developer for Intel® IPU (Infrastructure Processing Unit), programmable network devices managing and accelerating networking functions in a data center.
Fascinated by human cognition as well as machine learning, I’ve completed Master degrees in both Cognitive studies and Intelligent systems. I aim to use the strength of the machine to the benefit of humans…