AI Hackathon explores use of AI in everyday business life
How can the topic of artificial intelligence (AI) really be advanced in everyday business?
Used correctly, Artificial Intelligence and Data Science, for example, can open up a wide range of possibilities for making operational processes significantly more efficient. This is true of numerous industries and segments – but only a few subjects open up as much potential as the validation of autonomous driving functions. For good reason, MicroNova's Consulting division has for some time now been offering a comprehensive range of consulting and development services with a focus on Data Analysis and Machine Learning (ML); companies from the automotive sector receive valuable assistance for the challenges they face relating to the use of AI.
Because one thing is clear: AI must become widespread! But how can companies, in this case MicroNova, really push forward the issue operationally during ongoing everyday business and as such in addition to all orders and deadlines? How can employees be won over for an AI project? How can ideas be collected and implemented in such a way that they can be used in a customer project? And how can all this be done when everyone involved has little spare time available?
The answer is as simple as it is challenging: just do it! AI is not just something that can or should be approached strategically. Rather, there are a number of possible starting points, such as a project of individual people, as weekly meetings, or via dedicated brainstorming sessions. If an exchange of knowledge is desired and at the same time employees' precious time is to be used optimally to actually arrive at a functioning piece of software, there is another way: the hackathon.
In this event form, talents from different areas work on one specific problem for a limited period of time – and sometimes achieve amazing results. This is also the case at MicroNova, with its many brilliant experts. Since they all work at different sites, the latest hackathon from our AI Competence Team was a cross-border event, with colleagues from Mladá Boleslav in the Czech Republic, and Wolfsburg and Leonberg in Germany.
The topic of the hackathon was to test software including a graphical user interface. Due to time constraints, it was important that the tests were automated – because the software is developed agilely in a SCRUM project and therefore new releases have to be scrutinized regularly. Furthermore, the tests carried out should also be usable later for qualification of the software according to the ISO 26262 safety standard. The tools available to achieve this very often require adaptations to the software being tested before they can be used at all. In this case, however, not least for safety reasons, the software had to be tested in its original state.
A brainstorming session at the beginning of the hackathon revealed two possible solutions: the use of
- AI libraries to examine individual elements of the user interface or
- a freely available framework already developed for testing UIs.
Working groups followed up on both approaches. It soon became clear that the freely available framework would have had to be adapted to the use case at hand, which would require a lot of effort. The team therefore decided to develop its own EXAM library based on AI libraries. The working groups were divided again between development of the library, and testing and documentation; another group ensured the library's compatibility with the EXAM Test Case Generator.
The developed library can scan the user interface to identify the menus, buttons and text boxes that can be used to operate and test the software. This means the application being tested is operated automatically as if by a real user. In the next step, the library was used to implement the test specifications as automated test cases in EXAM. Using the Test Case Generator also saved time.
Ultimately and despite a tight deadline, the MicroNova team was able to achieve the hackathon's stated goal: The required library for implementation of the automated test cases was available. An extensive test of the library also proved that it was suitable for use in qualification tests in accordance with ISO 26262.
For us at MicroNova, this proved once again that AI can also be used in areas such as testing graphical user interfaces, and the solution can be quickly available with the right approach. Or to put it another way: If tackled correctly, progress on the subject of AI can be achieved in everyday life!
