Bugs in software are largely inevitable, but a new tool from Altran engineering and R&D services promises to reduce their occurrence. With the use of artificial intelligence (AI), Code Defect AI helps software developers predict the existence of bugs early in the development process when they are easier and less expensive to fix. According to the company, the tool applies machine learning to historical data in order to identify potential problem areas in the code and recommends tests to diagnose and fix any bugs. This speeds up development times and improves the quality of the software.

“It’s well known that software developers are under constant pressure to release code fast without compromising on quality,” said Walid Negm, Altran's chief innovation officer. “The reality, however, is that the software release cycle needs more than automation of assembly and delivery activities. It needs algorithms that can help make strategic judgments — especially as code gets more complex. Code Defect AI does exactly that.”

Using a variety of machine learning techniques, including random decision forests, support vector machines, multilayer perceptron and logistic regression, Code Defect AI extracts, preprocesses and labels the historical data to train the algorithm and put together a reliable decision model. Developers get a confidence score that predicts the code's compliance or risk of containing bugs. They can also use Code Defect AI to determine which feature(s) in the code should take precedence due to having higher prediction potential.

Code Defect AI is available on GitHub (owned by Microsoft) and on the Microsoft AI Lab portal. It also supports integration with other third-party analysis tools. It is scalable and can be hosted on-premise or on cloud computing platforms like Microsoft Azure. Learn more about Altran's productivity tools at Intelligent Automation.