JOB DESCRIPTION

The PhD will work towards analyzing explainable AI (XAI) techniques and designing explainable AI methods for diagnosing embedded AI models. They will focus on analyzing the state-of-the-art model-specific (e.g., integrated gradient) and model-free (e.g., counterfactual instances) XAI methods to identify robustness issues through causal analysis of relevant feature in model behavior. To that end, they will develop principled approaches and practical tools for diagnosing what knowledge the embedded ML model needs thus clearing the roadblock for robust AI. As part of the effort, they will design representation of knowledge of embedded AI and develop a reasoning engine to infer the unknowns of an embedded AI model. The expected results include 1) a human-in-the-loop knowledge extraction approach for describing AI knowledge in semantic concepts and for specifying required mechanisms in specific tasks; 2) a proof-of-concept of proposed diagnosis tool in the selected embedded AI domains.

JOB REQUIREMENTS

The successful PhD candidate should:

• Have an MSc in Computer Science, or a closely related field;
• Have demonstrated competencies in one or more of these domains: Crowd Computing, Machine Learning, Natural Language Processing, Data Science, Human Computer Interaction;
• Have strong interest and/or experience conducting empirical and theoretical research;
• Be self-motivated and enjoy research;
• Be a team player;
• Have good communication skills in Engligh.

Please refer the Job description for details