PhD Candidate: Explainable AI for Synthetic Media

Norwegian University of Science and Technology (NTNU); Gjøvik; On-site; NOK 493,000 - 667,000; Full time

Benefits offered by this job

International research environment

Collaboration opportunities

Pension benefits (SPK)

Norwegian language training

Job summary

NTNU invites applications for a PhD candidate position in machine learning at the Department of Computer Science (IDI), Gjøvik. The role is a three-year full-time doctoral appointment based at Gjøvik campus, reporting to the Unit Leader of Colorlab.

The project focuses on reliable methods for authenticating visual media, adapting pretrained representations, and ensuring explainability and robustness across datasets and real-world transformations.

Qualifications

• Master's degree in computer science, AI, ML, CV or signal processing with a substantial independent project.
• Strong knowledge of machine learning and deep learning with experimental research experience.
• Proficient in Python and a modern deep-learning framework (PyTorch or TensorFlow).

Responsibilities

• Develop ML methods to detect, localize, or characterize synthetic and manipulated visual content.
• Study representations learned by large pretrained visual or multimodal encoders.
• Design explanation methods linking predictions to actionable evidence.
• Evaluate robustness and transfer to unseen datasets and post-processing.

Skills

Machine learning

Deep learning

Python

PyTorch

TensorFlow

English communication

Education

Master's degree in Computer Science / AI / ML / CV

Tools

PyTorch

TensorFlow

Job description

NTNU invites applications for a PhD candidate position in machine learning at the Department of Computer Science (IDI), Gjøvik. The role is a three-year full-time doctoral appointment based at Gjøvik campus, reporting to the Unit Leader of Colorlab.

The project focuses on reliable methods for authenticating visual media, adapting pretrained representations, and ensuring explainability and robustness across datasets and real-world transformations.