Author : Apptest.ai

For the full slide deck, go to the link : https://www.slideshare.net/slideshow/embed_code/key/X0FucQ4YKsCyK

Among all mobile apps, the current error rate is believed to be at 15%. With a thousand new apps launching daily and a constant increase of mobile devices, there’s a need for a scalable solution to create and maintain high-quality apps hassle free.

Thanks to AI, exploratory testing is advancing and proving to detect mobile bugs at scale. We examined the five most common types of errors found through over 10,000 hours of AI-powered testing with actual samples. We introduce the challenges of each and explain how the heuristic or rule-based approach was not able to address the issues efficiently. The slide deck will cover topics such as broken element identification, and Z-order detection for layered views.

The five types of bugs we highlighted for demonstration are:

1. Broken link
2. NPE (Null Pointer Exception)
3. Index out of bound exception
4. Out of memory
5. Bad token exception

You can also look at how we leveraged deep learning methods such as rCNN and LSTM which enhance the coverage and reliability. The apptest.ai framework can be integrated into CI/CD environments for agile development.