About

I am a Postdoctoral researcher in Deep Learning and Computer Vision at the University of Oxford, working with Prof Ben Sheldon.

My research investigates how inductive biases derived from visual perception and physical signal formation shape representation learning in deep neural networks. I focus on embedding perceptually and physically grounded principles into model architectures, with the goal of improving robustness, interpretability, and generalisation under complex real-world conditions.

A central theme of my work is the design of convolutional and hybrid architectures that explicitly encode these priors. Rather than treating non-RGB sensing modalities or challenging environments as domain-specific applications, I use them as stress tests for modern deep learning, exposing systematic failure modes and motivating more principled architectural design.

I am particularly interested in multimodal and multi-spectral learning, where I treat the physical properties of sensed data as a source of inductive bias rather than nuisance variability. By grounding model design in the physics of image formation, my work develops representations that generalise across sensors, conditions, and domains.

Across these areas, I design and curate datasets as experimental instruments for analysing representation robustness, inductive bias, and generalisation under physically grounded real-world variability.

Current research themes:

Representation learning with explicit perceptual and physical inductive biases.
Architectures for robust semantic segmentation in complex visual environments.
Multimodal and multi-spectral learning under physical sensing constraints.
Domain-invariant representations under seasonal and cross-sensor shift.

Open to Collaboration:

I am always interested in discussing new research directions and potential collaborations around representation learning, inductive bias, and robust perception under real-world sensing constraints.

Dr Georgios Voulgaris

Recent News