School of Mathematics & Statistics

Dr Tiffany Vlaar

  • Lecturer in Applied Mathematics (Mathematics)

Biography

I am an ELLIS (European Laboratory for Learning and Intelligent Systems) Member. Prior to becoming a lecturer at the University of Glasgow, I was a postdoctoral researcher at Mila - Quebec AI Institute and McGill University in climate change AI and mathematics of deep learning. I obtained my Mathematics PhD from the University of Edinburgh, during which I was a Turing Enrichment student, and also have a background in physics (MSc, Perimeter Institute). I am co-organizing the International Congress of Mathematicians (ICM) in Glasgow in 2030.

 

Pronouns: she/her.

Research interests

See website for more info on research projects: tiffanyvlaar.github.io

Research groups

Publications

List by: Type | Date

Jump to: 2026 | 2025 | 2024 | 2022 | 2021 | 2019 | 2016
Number of items: 9.

2026

Frohlich, Tobias, Vlaar, Tiffany ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 and Thamsen, Lauritz ORCID logoORCID: https://orcid.org/0000-0003-3755-1503 (2026) Could Model Partitioning Make Federated Learning More Sustainable? In: 2nd International Workshop on Low Carbon Computing (loco 2026), Lancaster, UK, 10 -11 Sep 2026, (doi: 10.48550/arXiv.2608.14242)

2025

Hockerts, L., Stewart, Peter ORCID logoORCID: https://orcid.org/0000-0001-6338-6279 and Vlaar, T. ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 (2025) Camera-Trap Classification With Deep Learning Under Ground Truth Uncertainty. Tackling Climate Change with Machine Learning Workshop (NeurIPS 2025), San Diego, CA, USA, 30 Nov - 07 Dec 2025. (Accepted for Publication)

Zeise, Noam and Vlaar, Tiffany Joyce ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 (2025) BatchNorm Layers have an Outsized Effect on Adversarial Robustness. OPT2025: 17th Annual Workshop on Optimization for Machine Learning (NeurIPS 2025), San Diego, CA, USA, 30 Nov - 07 Dec 2025. (Accepted for Publication)

2024

Müller, Maximilian, Vlaar, Tiffany ORCID logoORCID: https://orcid.org/0000-0002-0885-2393, Rolnick, David and Hein, Matthias (2024) Normalization Layers Are All That Sharpness-Aware Minimization Needs. In: 37th Conference on Neural Information Processing Systems (NeurIPS 2023), New Orleans, Louisiana, USA, 10-16 December 2023,

2022

Vlaar, Tiffany ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 and Leimkuhler, Benedict (2022) Multirate Training of Neural Networks. In: 39th International Conference on Machine Learning (ICML2022), Baltimore, Maryland, USA, 17-23 July 2022, pp. 22342-22360.

Vlaar, Tiffany J. ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 and Frankle, Jonathan (2022) What Can Linear Interpolation of Neural Network Loss Landscapes Tell Us? In: 39th International Conference on Machine Learning (ICML2022), Baltimore, Maryland, USA, 17-23 July 2022, pp. 22325-22341.

2021

Leimkuhler, Benedict, Vlaar, Tiffany ORCID logoORCID: https://orcid.org/0000-0002-0885-2393, Pouchon, Timothée and Storkey, Amos (2021) Better Training using Weight-Constrained Stochastic Dynamics. In: 38th International Conference on Machine Learning (ICML2021), 18-24 July 2022, pp. 6200-6211.

2019

Leimkuhler, Benedict, Matthews, Charles and Vlaar, Tiffany ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 (2019) Partitioned integrators for thermodynamic parameterization of neural networks. Foundations of Data Science, 1(4), pp. 457-489. (doi: 10.3934/fods.2019019)

2016

Chojnacki, Leilee, Cook, Caleb Q., Dalidovich, Denis, Hayward Sierens, Lauren E., Lantagne-Hurtubise, Étienne, Melko, Roger G. and Vlaar, Tiffany J. ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 (2016) Shape dependence of two-cylinder Rényi entropies for free bosons on a lattice. Physical Review B, 94(16), 165136. (doi: 10.1103/PhysRevB.94.165136)

This list was generated on Sun Aug 23 17:32:01 2026 BST.
Jump to: Articles | Conference or Workshop Item | Conference Proceedings
Number of items: 9.

Articles

Leimkuhler, Benedict, Matthews, Charles and Vlaar, Tiffany ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 (2019) Partitioned integrators for thermodynamic parameterization of neural networks. Foundations of Data Science, 1(4), pp. 457-489. (doi: 10.3934/fods.2019019)

Chojnacki, Leilee, Cook, Caleb Q., Dalidovich, Denis, Hayward Sierens, Lauren E., Lantagne-Hurtubise, Étienne, Melko, Roger G. and Vlaar, Tiffany J. ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 (2016) Shape dependence of two-cylinder Rényi entropies for free bosons on a lattice. Physical Review B, 94(16), 165136. (doi: 10.1103/PhysRevB.94.165136)

Conference or Workshop Item

Hockerts, L., Stewart, Peter ORCID logoORCID: https://orcid.org/0000-0001-6338-6279 and Vlaar, T. ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 (2025) Camera-Trap Classification With Deep Learning Under Ground Truth Uncertainty. Tackling Climate Change with Machine Learning Workshop (NeurIPS 2025), San Diego, CA, USA, 30 Nov - 07 Dec 2025. (Accepted for Publication)

Zeise, Noam and Vlaar, Tiffany Joyce ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 (2025) BatchNorm Layers have an Outsized Effect on Adversarial Robustness. OPT2025: 17th Annual Workshop on Optimization for Machine Learning (NeurIPS 2025), San Diego, CA, USA, 30 Nov - 07 Dec 2025. (Accepted for Publication)

Conference Proceedings

Frohlich, Tobias, Vlaar, Tiffany ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 and Thamsen, Lauritz ORCID logoORCID: https://orcid.org/0000-0003-3755-1503 (2026) Could Model Partitioning Make Federated Learning More Sustainable? In: 2nd International Workshop on Low Carbon Computing (loco 2026), Lancaster, UK, 10 -11 Sep 2026, (doi: 10.48550/arXiv.2608.14242)

Müller, Maximilian, Vlaar, Tiffany ORCID logoORCID: https://orcid.org/0000-0002-0885-2393, Rolnick, David and Hein, Matthias (2024) Normalization Layers Are All That Sharpness-Aware Minimization Needs. In: 37th Conference on Neural Information Processing Systems (NeurIPS 2023), New Orleans, Louisiana, USA, 10-16 December 2023,

Vlaar, Tiffany ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 and Leimkuhler, Benedict (2022) Multirate Training of Neural Networks. In: 39th International Conference on Machine Learning (ICML2022), Baltimore, Maryland, USA, 17-23 July 2022, pp. 22342-22360.

Vlaar, Tiffany J. ORCID logoORCID: https://orcid.org/0000-0002-0885-2393 and Frankle, Jonathan (2022) What Can Linear Interpolation of Neural Network Loss Landscapes Tell Us? In: 39th International Conference on Machine Learning (ICML2022), Baltimore, Maryland, USA, 17-23 July 2022, pp. 22325-22341.

Leimkuhler, Benedict, Vlaar, Tiffany ORCID logoORCID: https://orcid.org/0000-0002-0885-2393, Pouchon, Timothée and Storkey, Amos (2021) Better Training using Weight-Constrained Stochastic Dynamics. In: 38th International Conference on Machine Learning (ICML2021), 18-24 July 2022, pp. 6200-6211.

This list was generated on Sun Aug 23 17:32:01 2026 BST.

Supervision

If you are interested in my research, please feel free to reach out to me to discuss PhD options at the University of Glasgow.
In your email please include your CV, university transcripts, and any dissertations/papers you may have written (but do not worry if you do not have any publications at this stage).

Opportunties:

  • PhD project within the IAPETUS DTP with Dr. Ellen Bowler (British Antarctic Survey), Prof. Phil Stephens (Durham University), Dr. Linus Ericsson (University of Glasgow), and Dr. Peter Fretwell (British Antarctic Survey) on From Wandering Albatrosses to Hedgehogs: Using AI and Citizen Science to Improve Biodiversity Monitoring under Ground Truth Uncertainty. 

    Applications will open later this year.
  • PhD project within NetGAIN with Prof. Phil Stephens (Durham University), Dr. Thomas Hodgson (CreditNature), Joel Sellors-Moore (National Trust), and Prof. Colin Torney (University of Glasgow) Towards increased confidence and public engagement in targeted AI-based frameworks for biodiversity monitoring. 

    Non-academic partners: CreditNature, National Trust. 

  • The University of Glasgow offers James McCune Smith scholarships for black UK domiciled students. More information: https://www.gla.ac.uk/scholarships/mccune-smith/

  • Feel free to contact me if you are interested in pursuing research on the interface of mathematics and deep learning.

Past supervision

  • Deploying deep learning for marine biodiversity monitoring with Dr. De Clippele (main, University of Glasgow) and Dr. Smith (University of Copenhagen).
  • UG Summer Research project on Monitoring vegetation patterns from drone imagery using machine learning, with Dr. Eizenhöfer (University of Glasgow).
  • Machine learning for camera trap image classification, EPSRC Vacation Project with Dr. Peter Stewart (University of Glasgow). Output: paper NeurIPS Tackling Climate Change with Machine Learning 2025 Workshop.
  • BatchNorm Layers have an Outsized Effect on Adversarial Robustness, Summer Project. Output: paper NeurIPS Optimization for Machine Learning 2025 Workshop.
  • AIMS Ghana MSc in Mathematical Sciences projects.
  • Effective nonlinearity and effective capacity of deep neural networks, Master Thesis Projects, with Dr. Pinson (main, Eindhoven University of Technology).
  • MSci project with industry co-supervisor Dr. Hodgson (CreditNature) on Simulating Scottish Ecosystems for Landscape Ecology Metrics.

PhD supervision:

  • A PhD project on Deep Learning Optimization and Design Choices for Marine Biodiversity Monitoring within the Leverhulme Programme for Doctoral Training in Ecological Data Science with Dr. Laurence De Clippele (University of Glasgow).
  • Fay Bennedik. Developing GPU-accelerated digital twins of ecological systems for population monitoring and scenario analyses project with Prof. McCrea (main, Lancaster University), Prof. Torney (University of Glasgow), Prof. Morales (University of Glasgow), and Prof. Husmeier (University of Glasgow) within the ExaGEO CDT. 
  • Antonella Marsella through the DiveIn CDT. Jointly supervised with Prof. Fani Deligianni (main, Computing Science ) and Dr. Stuart Grey (James Watt School of Engineering).
  • Sílvia da Silva Mourão through the DiveIn CDT. Supervised by: Dr. Paul R. Eizenhöfer (main, Geographical & Earth Sciences), Prof. Larissa Naylor (Geographical & Earth Sciences), and Dr. Zhiwei Gao (Engineering).
  • Michaelangelos Petrakis Rios. Dissecting deep neural networks. Co-supervised with Dr Linus Ericsson (Computing Science).

 

  • Frohlich, Tobias
    Carbon-Aware Model Partitioning for Federated Learning
  • Wu, Viktor
    The marine rewilding effect: Building a cost-effective framework for measuring Marine Net Gain

Teaching

Mathematics 1G: Introduction to Algebra, Geometry & Networks (MATHS1016).

Large-Scale Computing for Data Analytics (STATS5083), University of Glasgow.

Co-Lead, Widening Participation Summer School Mathematics, University of Glasgow.

Additional information

I was a board member of the One World Seminar on Mathematics of Machine Learning from 2022-25. Have a look at the talk recordings if you are interested.

I am currently an organizer of the Network on Mathematical Data Science for Materials Science. 

I am a Founding Member of the IMA's first Special Interest Group (SIG). Topic: Mathematics for Climate, Environment & Sustainability. 

I am passionate about increasing diversity in postgraduate research programmes in STEM. I was a Women in Machine Learning (WiML) board member and General Chair of the WiML workshop at NeurIPS 2024. I co-founded the Piscopia Initiative (piscopia.co.uk) to provide information about how to apply to PhDs and what doing a PhD is like. The events are particularly aimed at women and non-binary people in Mathematics, but anyone is welcome to attend. Check out the https://piscopia.co.uk/contacts/ section for info on related initiatives. 


Tiffany Vlaar