Key Skills

Job Description

Job Description Background Machine learning is increasingly used in aerospace for tasks that have traditionally relied on human experts or on fixed rule-based procedures. For such tasks, data-driven algorithms promise gains in speed, consistency and sensitivity. Before they can be used in practice however, they must be qualified, which means it has to be demonstrated with sufficient substantiation that the algorithm performs its intended function reliably within a defined set of conditions. The qualification approaches used in the sector today were developed for human operators and for conventional systems whose behaviour can be traced back to explicit requirements. They do not translate directly to algorithms whose behaviour is learned from data, and emerging guidance from aviation authorities names explainability as a key ingredient of trustworthy AI, while leaving open how it should be demonstrated in practice. Performance figures on a test set do not show why an algorithm reaches a decision, whether it relies on physically meaningful information, or how it behaves outside the data it was trained on. Explainable machine learning offers a wide range of methods to address this, from feature attribution and saliency methods, to example-based and concept-based explanations, but it is still an open question which of these can provide evidence that is strong enough to support qualification. This project therefore investigates how explainability methods can contribute to the qualification of machine learning algorithms for aerospace, starting from an use case, and provides a comparative assessment of several methods. Assignment The Assignment Will Include The Following Tasks A literature study on the qualification of machine learning algorithms in safety-critical sectors and on explainability methods for machine learning (e.g., attribution and saliency methods, example-based and concept-based explanations, interpretable models, uncertainty quantification), including the ways in which explanations themselves are evaluated. A use case provided by NLR, for instance automated defect detection, and identification of the substantiation needed to qualify the algorithm, translated into criteria that explanations should meet. Implementing several explainability methods for the use case and analysing them against these criteria, for example their faithfulness to the model, their robustness and their usefulness to domain experts. Assessing how the resulting explanations can serve as evidence in the qualification of the algorithm, and formulating recommendations for the use of explainability in qualification. Results The Final Outcomes Of This Assignment Will Be A structured assessment, as part of a literature study, of explainability methods for machine learning and their suitability for qualification in aerospace, together with a justified selection of the use case. An implementation and comparative evaluation of several explainability methods on the selected use case, with recommendations on how explainability can support the qualification of the algorithm. A technical report, either an MSc thesis following the guidelines of your faculty or an internship report, describing the approach, results, and conclusions of the work. Duration For an MSc thesis, the standard duration of a thesis at your faculty applies (typically 9 months). For an internship, the duration is 6 months. Profile You are an MSc student in Aerospace Engineering, Computer Science, Artificial Intelligence, Applied Mathematics or a similar master. You have a strong background in machine learning, preferably deep learning. Knowledge of explainable AI is a plus. You have experience with programming in Python and with machine learning libraries, for instance PyTorch or scikit-learn. You have an affinity with aerospace, inspection or certification, and are motivated to bring machine learning closer to qualified use in practice. What We Offer A challenging graduation project/internship in a high-tech result orientated work environment Weekly supervision and availability of the technical staff for support An internship allowance Working in an actual R&D project as part of the team Internship results to be used in the current and future projects About NLR Royal NLR has been the ambitious research organisation with the will to keep innovating for over 100 years. With that drive, we make the world of transportation safer, more sustainable, more efficient and more effective. We are on the threshold of breakthrough innovations. Plans and ideas start to move when these are fed with the right energy. Over 800 driven professionals work on research and innovation, from aircraft engineers to psychologists and from mathematicians to application experts. Our colleagues are happy to tell you what it's like to work at NLR. This assignment will be managed by the Avionics Systems and Maintenance Engineering (ASAM) department. Interested? Submit your application via the NLR portal, together with your (i) Motivation letter and (ii) CV and we will contact you as soon as possible. For more information about the assignment contact Alex Hurd ( [email protected] ) or Daniel Friesen ( [email protected] ).

Core Responsibilities

Review literature on qualifying machine learning in safety-critical sectors and on explainability methods, then select an aerospace use case and define criteria for useful explanations. Implement and compare explainability methods against those criteria, assess how their results can support algorithm qualification, and provide recommendations in a technical report.

Requirements

Applicants should be MSc students in Aerospace Engineering, Computer Science, Artificial Intelligence, Applied Mathematics, or a similar field, with a strong machine learning background, preferably in deep learning. Python and machine learning library experience are expected; explainable AI knowledge is a plus, along with an interest in aerospace, inspection, or certification.

Benefits

• Internship Allowance
• Weekly Supervision
• Technical Staff Support
• Participation in an R&D Project