Calgary Machine Learning Lab

The Calgary Machine Learning Lab is a research group led by Yani Ioannou within the Schulich School of Engineering at the University of Calgary. Our research is driven by the overarching goal of advancing efficient, trustworthy, and accessible Artificial Intelligence.

The Calgary ML Lab outside the ICT building at the University of Calgary — Lab photo from May 2026 outside the ICT building at the University of Calgary.

Central to our work is the concept of sparse neural network training and inference, which we pursue with four key motivations:

Democratize AI: removing redundant computation, making state-of-the-art models accessible to all
Sustainable and Trustworthy AI: by fundamentally reducing the carbon footprint of these models while rigorously auditing how compression impacts algorithmic bias, we are working to ensure that real-world deployment of next-generation AI is both sustainable and safer
Learning Structure in Neural Networks: automatically learn neural network topologies tailored for novel data domains
Understanding the Mechanics of Neural Network Training: sparse neural network training provides a unique theoretical lens to better understand the underlying principles of neural network training, and its remarkable effectiveness.

news

Sep 29, 2026	Yani Ioannou participated in the Wallenberg Advanced Scientific Forum, Autumn 2026: Symmetries in Neural Networks (archived excerpt), held September 29 to October 2, 2026 at Rånäs Castle, Sweden. The invitation-only forum brings together researchers in mathematics, computer science and engineering to work on how symmetries in data, weight space and training dynamics affect training and generalization.
Sep 25, 2026	CML is featured in the Gauntlet — the University of Calgary’s independent student-run newspaper — where we discuss the challenges of democratizing AI and the opportunities for students interested in AI research to help address them (archived copy).
Aug 18, 2026	Yufan Feng, has been awarded an Alberta Graduate Excellence Scholarship (AGES) (archived) to support her PhD studies, and Mike Lasby has been awarded the University of Calgary Silver Anniversary Graduate Fellowship (archived). Both scholarships are based on outstanding academic achievement.

latest blog posts

Jul 27, 2026	SparseOpt: Why Normalization Hurts Sparse Training (And How to Fix It)
Jul 14, 2025	Sparse Training from Random Initialization: Aligning Lottery Ticket Masks using Weight Symmetry
Mar 31, 2025	Beyond Compression: How Knowledge Distillation Impacts Fairness and Bias in AI Models

selected publications

NeurIPS

The Surprising Effectiveness of Deleting Weights in LLM Reasoning and Adaptation

Jack Lu, Zhenbang Yang, Mike Lasby, Tejas Pote, Yani Ioannou, and Mengye Ren

In The Fortieth Annual Conference on Neural Information Processing Systems, Dec 2026

Bib OpenReview

@inproceedings{lu2026surprising,
  title = {The Surprising Effectiveness of Deleting Weights in {LLM} Reasoning and Adaptation},
  author = {Lu, Jack and Yang, Zhenbang and Lasby, Mike and Pote, Tejas and Ioannou, Yani and Ren, Mengye},
  booktitle = {The Fortieth Annual Conference on Neural Information Processing Systems},
  month = dec,
  year = {2026},
}

ICML

SparseOpt: Addressing Normalization-Induced Gradient Skew in Sparse Training

Mohammed Adnan, Rohan Jain, Tom Jacobs, Ekansh Sharma, Rahul G. Krishnan, Rebekka Burkholz, and Yani Ioannou

In Proceedings of the 43rd International Conference on Machine Learning (ICML), Jul 2026

arXiv Bib PDF OpenReview Website

@inproceedings{adnan2026sparseopt,
  title = {{SparseOpt: Addressing Normalization-Induced Gradient Skew in Sparse Training}},
  author = {Adnan, Mohammed and Jain, Rohan and Jacobs, Tom and Sharma, Ekansh and Krishnan, Rahul G. and Burkholz, Rebekka and Ioannou, Yani},
  booktitle = {Proceedings of the 43rd International Conference on Machine Learning (ICML)},
  venue = {Seoul, South Korea},
  eventdate = {2026-07-05/2026-07-11},
  month = jul,
  year = {2026},
  eprinttype = {arXiv},
  eprint = {2605.27541},
  arxivid = {2605.27541},
  archiveprefix = {arXiv},
  primaryclass = {cs.LG},
  url = {https://openreview.net/forum?id=o9JP7N8YO9},
}

ICLR

REAP the Experts: Why Pruning Prevails for One-Shot MoE compression

Mike Lasby, Ivan Lazarevich, Nish Sinnadurai, Sean Lie, Yani Ioannou, and Vithursan Thangarasa

In International Conference on Learning Representations (ICLR), Apr 2026

arXiv Bib PDF Blog Code OpenReview Poster

@inproceedings{lasby2025reapexpertspruningprevails,
  title = {REAP the Experts: Why Pruning Prevails for One-Shot MoE compression},
  author = {Lasby, Mike and Lazarevich, Ivan and Sinnadurai, Nish and Lie, Sean and Ioannou, Yani and Thangarasa, Vithursan},
  year = {2026},
  eprinttype = {arXiv},
  eprint = {2510.13999},
  archiveprefix = {arXiv},
  arxivid = {2510.13999},
  primaryclass = {cs.LG},
  month = apr,
  booktitle = {{International Conference on Learning Representations (ICLR)}},
  venue = {{Rio de Janeiro, Brazil}},
  eventdate = {2026-04-23/2026-04-27},
}

ICML

Sparse Training from Random Initialization: Aligning Lottery Ticket Masks using Weight Symmetry

Mohammed Adnan, Rohan Jain, Ekansh Sharma, Rahul Krishnan, and Yani Ioannou

In Proceedings of the 42nd International Conference on Machine Learning (ICML), Jul 2025

arXiv Bib Video Blog Code OpenReview Poster Website

@inproceedings{mohammed2025sparsetraining,
  author = {Adnan, Mohammed and Jain, Rohan and Sharma, Ekansh and Krishnan, Rahul and Ioannou, Yani},
  title = {Sparse Training from Random Initialization: Aligning Lottery Ticket Masks using Weight Symmetry},
  booktitle = {Proceedings of the 42nd International Conference on Machine Learning (ICML)},
  pages = {477--498},
  editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry},
  volume = {267},
  series = {Proceedings of Machine Learning Research},
  publisher = {PMLR},
  month = jul,
  year = {2025},
  url = {https://proceedings.mlr.press/v267/adnan25a.html},
  arxivid = {2505.05143},
  eprint = {2505.05143},
  eprinttype = {arXiv},
  venue = {{Vancouver, BC, Canada}},
  eventdate = {2025-07-13/2025-07-19},
}

TMLR

What is Left After Distillation? How Knowledge Transfer Impacts Fairness and Bias

Aida Mohammadshahi, and Yani Ioannou

Transactions on Machine Learning Research (TMLR), Mar 2025

arXiv Bib PDF Video Blog OpenReview Poster Slides

@article{mohammadshahi2025leftafterdistillation,
  author = {Mohammadshahi, Aida and Ioannou, Yani},
  title = {What is Left After Distillation? How Knowledge Transfer Impacts Fairness and Bias},
  journal = {Transactions on Machine Learning Research (TMLR)},
  year = {2025},
  month = mar,
  arxivid = {2410.08407},
  eprint = {2410.08407},
  eprinttype = {arXiv},
}

NeurIPS

Navigating Extremes: Dynamic Sparsity in Large Output Spaces

Nasib Ullah, Erik Schultheis, Mike Lasby, Yani Ioannou, and Rohit Babbar

In 38th Annual Conference Neural Information Processing Systems (NeurIPS), Dec 2024

arXiv Bib PDF Code OpenReview Website

@inproceedings{nasib2024navigating,
  author = {Ullah, Nasib and Schultheis, Erik and Lasby, Mike and Ioannou, Yani and Babbar, Rohit},
  booktitle = {{38th Annual Conference Neural Information Processing Systems (NeurIPS)}},
  venue = {{Vancouver, BC, Canada}},
  eventdate = {2024-12-10/2024-12-16},
  title = {Navigating Extremes: Dynamic Sparsity in Large Output Spaces},
  month = dec,
  year = {2024},
  arxivid = {2411.03171},
  eprint = {2411.03171},
  eprinttype = {arXiv},
  url = {https://nips.cc/virtual/2024/poster/95193},
}

ICLR

Dynamic Sparse Training with Structured Sparsity

Mike Lasby, Anna Golubeva, Utku Evci, Mihai Nica, and Yani Ioannou

In International Conference on Learning Representations (ICLR), May 2024

arXiv Bib PDF Blog Code OpenReview Poster Slides Website

@inproceedings{lasby2024srigl,
  author = {Lasby, Mike and Golubeva, Anna and Evci, Utku and Nica, Mihai and Ioannou, Yani},
  booktitle = {{International Conference on Learning Representations (ICLR)}},
  venue = {{Vienna, Austria}},
  eventdate = {2024-05-07/2024-05-11},
  title = {Dynamic Sparse Training with Structured Sparsity},
  month = may,
  year = {2024},
  arxivid = {2305.02299},
  eprint = {2305.02299},
  eprinttype = {arXiv},
  url = {https://iclr.cc/virtual/2024/poster/17975},
  dimensions = {1157762278},
}