Dr Linus Ericsson

  • Lecturer in Artificial Intelligence / Machine Learning (School of Computing Science)

Biography

I am a lecturer in AI & ML in the IDA section of the School of Computing Science. My research explores AI/ML methods to learn transferable representations of data, building efficient neural networks, and adapting these models across data shifts, all to help people solve problems reliably across different scenarios. Specific interests include representation learning, multi-modal learning, robustness, and automated machine learning (AutoML).

Previously, I was a postdoctoral researcher at the University of Edinburgh, where I worked with Elliot J. Crowley on AutoML and efficient neural network architectures. I hold a PhD from the University of Edinburgh, where my thesis, Self-Supervised Learning for Transferable Representations, was supervised by Tim Hospedales.

Research interests

Research groups

  • Information, Data & Analysis Section

Publications

Prior publications

Other

Qin, S., Auras, A., Cohen, S.B., Crowley, E.J., Moeller, M., Ericsson, L., Lukasik, J. (2025) ONNX-NET: TOWARDS UNIVERSAL REPRESENTATIONS AND INSTANT PERFORMANCE PREDICTION FOR NEURAL ARCHITECTURES Arxiv Scopus - Elsevier. ISSN 23318422 (doi: 10.48550/arXiv.2510.04938)

Qin, S., Kadlecová, G., Pilát, M., Cohen, S.B., Neruda, R., Crowley, E.J., Lukasik, J., Ericsson, L. (2025) Transferrable Surrogates in Expressive Neural Architecture Search Spaces Arxiv Scopus - Elsevier. ISSN 23318422 (doi: 10.48550/arXiv.2504.12971)

Yang, C., Chen, Z., Espinosa, M., Ericsson, L., Wang, Z., Liu, J., Crowley, E.J. (2024) PlainMamba: Improving Non-Hierarchical Mamba in Visual Recognition Arxiv Scopus - Elsevier. ISSN 23318422 (doi: 10.48550/arXiv.2403.17695)

Espinosa, M., Yang, C., Ericsson, L., McDonagh, S., Crowley, E.J. (2024) There is no SAMantics! Exploring SAM as a Backbone for Visual Understanding Tasks Arxiv Scopus - Elsevier. ISSN 23318422 (doi: 10.48550/arXiv.2411.15288)

Ericsson, L., Espinosa, M., Yang, C., Antoniou, A., Storkey, A., Cohen, S.B., McDonagh, S., Crowley, E.J. (2024) einspace: Searching for Neural Architectures from Fundamental Operations Arxiv Scopus - Elsevier. ISSN 23318422 (doi: 10.48550/arXiv.2405.20838)

Ericsson, L., Li, D., Hospedales, T.M. (2023) Better Practices for Domain Adaptation Arxiv Scopus - Elsevier. ISSN 23318422 (doi: 10.48550/arXiv.2309.03879)

Dutt, R., Ericsson, L., Sanchez, P., Tsaftaris, S.A., Hospedales, T. (2023) Parameter-Efficient Fine-Tuning for Medical Image Analysis: The Missed Opportunity Arxiv Scopus - Elsevier. ISSN 23318422 (doi: 10.48550/arXiv.2305.08252)

Eastwood, C., von Kügelgen, J., Ericsson, L., Bouchacourt, D., Vincent, P., Schölkopf, B., Ibrahim, M. (2023) Self-Supervised Disentanglement by Leveraging Structure in Data Augmentations Arxiv Scopus - Elsevier. ISSN 23318422 (doi: 10.48550/arXiv.2311.08815)

Dong, N., Ericsson, L., Yang, Y., Leonardis, A., McDonagh, S. (2022) Label-Efficient Object Detection via Region Proposal Network Pre-Training Arxiv Scopus - Elsevier. ISSN 23318422 (doi: 10.48550/arXiv.2211.09022)

Ericsson, L., Gouk, H., Loy, C.C., Hospedales, T.M. (2021) Self-supervised representation learning: Introduction, advances and challenges Arxiv Scopus - Elsevier. ISSN 23318422 (doi: 10.48550/arxiv.2110.09327)

Thesis

Linus Ericsson (2024) Self-supervised learning for transferable representations Linus Ericsson. (doi: 10.7488/ERA/4162)

Article

Dong, N., Ericsson, L., Yang, Y., Leonardis, A., McDonagh, S. (2024) Label-efficient object detection via region proposal network pre-training Neurocomputing Scopus - Elsevier. ISSN 09252312 18728286 (doi: 10.1016/j.neucom.2024.127376)

Linus Ericsson, Henry Gouk, Chen Change Loy, Timothy M. Hospedales (2022) Self-Supervised Representation Learning: Introduction, Advances, and Challenges IEEE Signal Processing Magazine Linus Ericsson. ISSN 1558-0792 (doi: 10.1109/msp.2021.3134634)

Conference Proceedings

Dutt, R., Ericsson, L., Sanchez, P., Tsaftaris, S.A., Hospedales, T. (2024) Parameter-Efficient Fine-Tuning for Medical Image Analysis: The Missed Opportunity Proceedings of Machine Learning Research Scopus - Elsevier. ISSN 26403498

Ericsson, L., Espinosa, M., Yang, C., Antoniou, A., Storkey, A., Cohen, S.B., McDonagh, S., Crowley, E.J. (2024) einspace: Searching for Neural Architectures from Fundamental Operations Advances in Neural Information Processing Systems Scopus - Elsevier. ISSN 10495258

Ericsson, L., Li, D., Hospedales, T.M. (2023) Better Practices for Domain Adaptation Proceedings of Machine Learning Research Scopus - Elsevier. ISSN 26403498

Linus Ericsson, Henry Gouk, Timothy M. Hospedales (2022) Why Do Self-Supervised Models Transfer? On the Impact of Invariance on Downstream Tasks BMVC Linus Ericsson. (doi: 10.48550/ARXIV.2111.11398)

Ericsson, L., Gouk, H., Hospedales, T.M. (2022) Why Do Self-Supervised Models Transfer? On the Impact of Invariance on Downstream Tasks Bmvc 2022 33rd British Machine Vision Conference Proceedings Scopus - Elsevier.

Linus Ericsson, Henry Gouk, Timothy M. Hospedales (2021) How Well Do Self-Supervised Models Transfer? CVPR Linus Ericsson. (doi: 10.48550/ARXIV.2011.13377)

Ericsson, L., Gouk, H., Hospedales, T.M. (2021) How Well Do Self-Supervised Models Transfer? Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition Scopus - Elsevier. ISBN 9781665445092 ISSN 10636919 (doi: 10.1109/CVPR46437.2021.00537)

Supervision

I am currently seeking PhD students! There are several routes for application, including available scholarships within the School of Computing Science and the DiveIn CDT. Please take a look at my publications to see if our research interests align, and feel free to get in touch by email if you’d like to work with me.

Teaching

Courses

  • Computing Science Level 5 Programming for AI

Linus Ericsson