PhD in Fine-Grained Visual Understanding

Overview

As a PhD student, you will be embedded in the research group Computational Imaging and Deep Learning (CIDL), part of the Leiden Institute of Advanced Computer Science (LIACS). The project is part of the NXTGen High-tech project, focusing on new high-tech equipment and data-driven techniques for future industries. We collaborate with GREEFA and Fruitmasters on fruit sorting automation. The project emphasises developing and analysing general Computer Vision methods using real-world image data.

The PhD will address open research questions in Computer Vision related to fine-grained visual understanding and anomaly detection under limited, weak, or imperfect supervision. The precise direction will be defined with the candidate and may evolve during the PhD. Example directions include:

• Anomaly, novelty, and out-of-distribution detection in visual data
• Fine-grained visual understanding for distinguishing subtle irregularities
• Learning visual representations with limited, weak, or noisy supervision
• Adapting, specializing, or probing large pre-trained models for domain-specific visual understanding
• Self-supervised and representation learning for images and videos
• Robustness and generalisation in real-world vision systems
• Interpretability and explainability of anomaly detection and visual decision-making

As a candidate, you must have a strong background in machine learning and computer vision, and excellent programming skills. You should also have an understanding and interest in real-world applications.

• Conduct original and novel research in Computer Vision and Machine Learning
• Develop and analyse novel deep-learning methods for visual representation learning
• Publish and present research at leading international conferences and journals
• Contribute to teaching activities within LIACS
• Write and defend a PhD dissertation
• Engage in collaborations with colleagues across domains

The Leiden Faculty of Science is a world-class faculty with a dynamic international environment. The research spans artificial intelligence, computer science, mathematics, astronomy, physics, chemistry, bio-pharmaceutical sciences, biology, and environmental sciences. The faculty is located at Leiden’s Bio Science Park. For more information, see the Leiden Science website.

The Leiden Institute of Advanced Computer Science (LIACS) is the AI and Computer Science Institute of the Faculty of Science. We offer courses at Bachelor and Master levels in AI, Computer Science, and Data Science, and interdisciplinary areas. We strive for excellence in a caring institute with a clear career path for young scientists. For more information about LIACS, see the LIACS and Leiden University websites.

• A MSc degree in Computer Science, Artificial Intelligence, or a related field
• Strong background in Computer Vision and/or Machine Learning
• Motivation to perform foundational research and apply methods to real-world problems
• Motivation to work in a team towards real-world applications
• Creative, independent, and communicative team player
• Experience with writing scientific manuscripts and good academic writing skills
• Experience with deep learning frameworks and excellent programming skills (preferably Python)
• Excellent proficiency in English (oral and written)
• Flexibility regarding on-site work with diverse project partners

All PhD students are embedded in the Leiden University Graduate School of Science. The graduate school offers PhD training courses at three levels and advanced courses to deepen scientific knowledge. We also offer:

• An employment contract for an initial one year, with extension to three more years upon positive evaluation; contract falls under the CLA of Dutch Universities
• A salary between €3,059 and €3,881 gross per month, full-time (38 hours) Pay scale P
• Holiday allowance (8%), year-end bonus (8.3%), and ABP pension
• Full reimbursement of public transport commuting costs
• Flexible working hours with leave entitlements and options for extra leave or sabbatical
• Options for secondary employment conditions, such as sabbatical or paid parental leave
• Hybrid working if applicable in the…