positions

Task: Diagnosing Embedded AI Models (WP3)

Host institution: TU Delft

Country: Netherlands

Supervisor: Dr. J. Yang [TU Delft]

Co-supervisors: Dr. P. Mauroux [Fribourg]; Dr. M. Petković [Philips]

Objectives: 1) To develop principled approaches and practical tools for diagnosing what knowledge the embedded ML model needs thus clearing the roadblock for robust AI; 2) To obtain representation of knowledge of Embedded AI; 3) To develop a reasoning engine to infer the unknowns of an Embedded AI model.

Expected Results: 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) develop proof-of-concept of proposed diagnosis tool in the selected Embedded AI domains.

PhD enrolment: Doctoral School of TU Delft

Planned secondments: 

• Philips (4 months, M16-M19): interfaces for human description of model behaviours and required mechanisms in semantic concepts, with Dr. M. Petković (KPI: joint paper)

• UNIFR (4 months, M25-M28): demo and code release of abductive reasoning engine for model diagnosis (KPI: joint paper with demo)

Candidate profile: computer science, artificial intelligence, data science, electrical engineering, human-computer interaction (in order of preference)

Desirable skills/interests: machine learning, probabilistic reasoning, data analytics, interface design, system engineering (the applicant should be proficient in at least one or two of the skills)

Application Deadline: February 14, 2025, AoE

Descriptions of all the 18 DC Positions

Submit Your Application HERE