Fabiano Dalpiaz

Date: 15:15 – 16:00, Tuesday, 10.03.2020
Location: Minnaert – 2.02

Title: Requirements Classification with Interpretable Machine Learning and Dependency Parsing
Abstract: Requirements classification is a traditional application of machine learning (ML) for handling large requirements datasets. A prime example of a classification problem applied to requirements engineering (RE) is the distinction between functional and non-functional requirements. State-of-the-art classifiers build their effectiveness on a large set of word features like text n-grams or POS n-grams, which do not fully capture the essence of a requirement. As a result, it is arduous for human analysts to interpret the classification results by exploring the classifier’s inner workings. In this talk, which is based on a paper presented at the 2019 IEEE Requirements Engineering conference, we propose the use of more general linguistic features, such as dependency types, for the construction of interpretable ML classifiers for RE. We build our “interpretable” feature set, which contains only 17 features, using interpretable ML tools that allow us to analyze the inner workings of black-box ML classifiers. We will conclude the talk with reflections on the nature of the classification task at hand that derive from the difficulty of obtaining consistent gold standards.