TestCon Europe 2020

HYBRID EDITION

October 13-15

Vilnius and Online

Arnika Hryszko and Jarosław Hryszko

Jaroslaw: Advisor to Polish Testing Board, AI Expert, Quality Evangelist & Books Author; Arnika: “Exploratory Tester of the Year” in the “Testing Heroes 2018” Plebiscite

hryszko.pl, Poland

Arnika Hryszko and Jarosław Hryszko

Jaroslaw: Advisor to Polish Testing Board, AI Expert, Quality Evangelist & Books Author; Arnika: “Exploratory Tester of the Year” in the “Testing Heroes 2018” Plebiscite

hryszko.pl, Poland

Biography

Jaroslaw used to say about himself as a quality assurance hipster in IT – he tested software before it became fashionable. He specializes in the use of the latest technologies there, especially artificial intelligence mechanisms. He willingly shares his knowledge as a trainer and consultant. He also has extensive experience in management, including commanding military units. He studied (successfully) cybernetics, electronics and nuclear power.

Jarek is the author and co-author of a series of books, publications and presentations on nuclear & computer system safety, artificial intelligence and quality assurance in computer science, including “Software testing in practice”, “Requirements engineering in practice”, “Software engineering: Challenges and Solutions” and “Software Engineering: Improving practice through research”.

Since 2012, he has been conducting research on artificial intelligence to ensure quality in IT systems. He is a well-known conference speaker at more than 20 scientific and professional conferences.

Jarek is an advisor to the Polish Testing Board and a member of the Internet of Things Working Group at the Polish Ministry of Digital Affairs. He has also co-created a popularization program for the future Polish nuclear power industry.

When he is not in front of a computer screen, you can find him in a garage, at a shooting range, somewhere far away in the mountains, in the middle of the ocean or… at the Institute of Computer Science and Computer Mathematics of the Jagiellonian University.

Arnika is software tester with over 10 years of experience in the field of quality assurance. “Exploratory Tester of the Year” in the “Testing Heroes 2018” plebiscite. During her professional career, she gained extensive experience, ranging from manual tests, through websites, performance tests, mobile applications, data warehouses, business intelligence, embedded systems, advanced automation, to using machine learning in testing.

Co-author of publications and presentations on quality assurance in IT, including “Requirements Engineering in practice”. She is also a conference speaker, presenting at over 20 industry conferences.

Workshop

Predicting the Future – Machine Learning for Testers

Time & Date

9:00, 13 October

Place

Online

Language

English

Abstract

Let’s learn how to use machine learning to predict defect-prone areas in the code just being written! Why? To better plan your quality assurance actions! You don’t have to be artificial intelligence master – during the the workshop we will learn how to use the KNIME platform, which makes using machine learning very easy.

The benefits of knowing which elements of developed software will be most defective are rather obvious – for example – having such knowledge, programmers can apply  more unit tests in such areas, while testers can prioritize tests related with these elements.

During the workshop we will use the KNIME, which uses graphical, user-friendly interface. Complex data analysis, including that based on artificial intelligence mechanisms, in KNIME is done just by putting together diagrams from “blocks”, where each “block” is responsible for a small element of data analysis. In the introduction to the workshop we will learn to use this platform.

During the workshop we will prepare training data describing the occurrence of defects in the historical software version. Then we will create a predictive model by training the selected machine learning mechanism. Next, we will use created model to indicate possibly defective areas on data describing the new software. Finally, with a clever trick, we will check how accurate such defect-proness prediction was.

Agenda

Part 1: Introduction – what’s all this for?

  • What is machine learning
  • How and when it all started
  • How it can be used in software development
  • How it can be applied for improving quality

Part 2: Basics of working with KNIME

  • History of KNIME creation
  • Basic KNIME functions
  • KNIME plug-ind and addons
  • Simple workflow creation

Part 3: Preparation of the training data from the old software version

  • Importing data from different sources
  • Cleaning and preparing the data
  • Workflow properties
  • Merging the data
  • Using output functions to present the outcome

Part 4: Machine learning and prediction for new code

  • Different machine learning algorithms
  • Using learner block
  • Using predictor block

Part 5: Is it correct? – verifying our method

  • How to verify correctness of the outcome
  • Comparing different models and algorithms

Part 6: What’s next? Methodology based on prediction

  • How this prediction can be used on a daily basis
  • How to incorporate defect prediction into development process

Objectives

The goal of this workshop is to introduce the participants with machine learning, Knime tool and its use in predicting defect-prone areas in the code. Also, how to apply this knowledge to provide better quality.

Target audience

People interested in the practical use of Artificial Intelligence mechanisms in planning software quality assurance processes. Testers, programmers, test managers. Knowledge of the topics and technical details of machine learning is not required.

Technical requirements

  • To participate in the workshop you will need a laptop, an installed KNIME framework with DePress and sample data.

  • The laptop should run decent versions of Windows, MacOS or Linux.
  • KNIME for the above operating systems can be downloaded free of charge from the website
  • You can install DePress add-on by following the instructions
  • In the case of data,we will use one of the Open Source projects data, available here 

« Back