|
|
AI Solutions and Data Harvesting in Testing |
The problem of triaging failing tests in the testing context can be time-consuming and disruptive. The proposed solution is to leverage machine learning and historical data to classify regression failures. This classification will be presented as a recommendation system, providing the test regression owner with a "Failure Reason/Area" along with a confidence/probability level. Our team implemented a Proof of Concept (PoC) using this approach and achieved promising results. We will discuss the details of the PoC, its outcomes, and the next steps to be taken. Additionally, we will address the challenges encountered during the implementation. Furthermore, we will explore the potential generalization of this data-based technique for other purposes beyond test triaging. By harnessing the power of AI and data harvesting, we can uncover new applications and benefits in various domains. Ganit Prisant Henkin Ganit has an extensive background as a software engineer, specializing in embedded Real-Time environments, with over 25 years of experience. She began her career at Freescale (formerly Motorola) Semiconductors and currently holds the position of firmware micro-architect and developer for Intel® IPU (Infrastructure Processing Unit). In this role, she works on programmable network devices that manage and accelerate networking functions within a data center. Driven by a fascination for human cognition and machine learning, Ganit pursued Master's degrees in both Cognitive Studies and Intelligent Systems. Her objective is to leverage the power of machines for the benefit of humanity, seeking to bridge the gap between human intelligence and machine capabilities.
|
|
|
|
|
|
|
