Publications

Showing all 56 publications

2026

NeurIPS

AsyncMesh: Fully Asynchronous Optimization for Data and Pipeline Parallelism

T. Ajanthan, S. Ramasinghe, G. Avraham, H. Mohaghegh Dolatabadi, C. P. Hewa Koneputugodage, V. Shevchenko, Y. Zuo, A. Long

PDFarXivCode
BibTeX
@inproceedings{ajanthan_asyncmesh_arxiv26,
  title={{AsyncMesh: Fully Asynchronous Optimization for Data and Pipeline Parallelism}},
  author={Ajanthan, Thalaiyasingam and Ramasinghe, Sameera and Avraham, Gil and Mohaghegh Dolatabadi, Hadi and P Hewa Koneputugodage, Chamin and Shevchenko, Violetta and Zuo, Yan and Long, Alexander},
  year={2026},
  booktitle={NeurIPS}
}

Communication-Efficient LLM Adaptation over Decentralized GPU Meshes

S. Ramasinghe, S. Siriwardhana, T. Ajanthan, H. Mohaghegh Dolatabadi, C. P. Hewa Koneputugodage, G. Avraham, V. Shevchenko, J. Snewin, K. Pajak, H. Xi, A. Long

PDFarXiv
BibTeX
@inproceedings{ramasinghe_llm_adaptation_arxiv26,
  title={{Communication-Efficient LLM Adaptation over Decentralized GPU Meshes}},
  author={Ramasinghe, Sameera and Siriwardhana, Shamane and Ajanthan, Thalaiyasingam and Mohaghegh Dolatabadi, Hadi and P Hewa Koneputugodage, Chamin and Avraham, Gil and Shevchenko, Violetta and Snewin, James and Pajak, Karol and Xi, Harry and Long, Alexander},
  year={2026},
  booktitle={NeurIPS}
}

NuMuon: Nuclear-Norm-Constrained Muon for Compressible LLM Training

H. Mohaghegh Dolatabadi, T. Ajanthan, S. Ramasinghe, C. P. Hewa Koneputugodage, S. Siriwardhana, V. Shevchenko, K. Pajak, J. Snewin, G. Avraham, A. Long

PDFarXiv
BibTeX
@inproceedings{dolatabadi_numuon_arxiv26,
  title={{NuMuon: Nuclear-Norm-Constrained Muon for Compressible LLM Training}},
  author={Mohaghegh Dolatabadi, Hadi and Ajanthan, Thalaiyasingam and Ramasinghe, Sameera and P Hewa Koneputugodage, Chamin and Siriwardhana, Shamane and Shevchenko, Violetta and Pajak, Karol and Snewin, James and Avraham, Gil and Long, Alexander},
  year={2026},
  booktitle={NeurIPS}
}

ICML

Factored Gossip DiLoCo: Reducing Blocking Communication in DiLoCo

C. Hewa Koneputugodage, T. Ajanthan, S. Ramasinghe, H. Mohaghegh Dolatabadi, S. Siriwardhana, G. Avraham, V. Shevchenko, K. Pajak, J. Snewin, A. Long

PDFarXivPost
BibTeX
@inproceedings{koneputugodage_factored_gossip_icml26,
  title={{Factored Gossip DiLoCo: Reducing Blocking Communication in DiLoCo}},
  author={Hewa Koneputugodage, Chamin and Ajanthan, Thalaiyasingam and Ramasinghe, Sameera and Mohaghegh Dolatabadi, Hadi and Siriwardhana, Shamane and Avraham, Gil and Shevchenko, Violetta and Pajak, Karol and Snewin, James and Long, Alexander},
  year={2026},
  booktitle={ICML}
}

ICLR

Taming Curvature: Architecture Warm-up for Stable Transformer Training

S. Ramasinghe, T. Ajanthan, H. Mohaghegh Dolatabadi, C. P. Hewa Koneputugodage, G. Avraham, V. Shevchenko, Y. Zuo, K. Pajak, A. Long

PDFarXiv
BibTeX
@inproceedings{ramasinghe_archwp_iclr26,
  title={{Taming Curvature: Architecture Warm-up for Stable Transformer Training}},
  author={Ramasinghe, Sameera and Ajanthan, Thalaiyasingam and Mohaghegh Dolatabadi, Hadi and P Hewa Koneputugodage, Chamin and Avraham, Gil and Shevchenko, Violetta and Zuo, Yan and Pajak, Karol and Long, Alexander},
  year={2026},
  booktitle={ICLR}
}

Preprints

Agora: Collective and Permissionless Internet-Scale Pretraining of Large Language Models

G. Avraham*, V. Shevchenko*, H. Mohaghegh Dolatabadi*, K. Pajak*, J. Snewin, H. Xi, R. O'Donnell, T. Ajanthan, S. Ramasinghe, C. Hewa Koneputugodage, S. Siriwardhana, A. Long

PDFarXivCode
BibTeX
@article{avraham_agora_arxiv26,
  title={{Agora: Collective and Permissionless Internet-Scale Pretraining of Large Language Models}},
  author={Avraham, Gil and Shevchenko, Violetta and Mohaghegh Dolatabadi, Hadi and Pajak, Karol and Snewin, James and Xi, Harry and O'Donnell, Rodney and Ajanthan, Thalaiyasingam and Ramasinghe, Sameera and Hewa Koneputugodage, Chamin and Siriwardhana, Shamane and Long, Alexander},
  year={2026},
  journal={arXiv},
  eprint={2607.13332},
  archiveprefix={arXiv},
  url={https://arxiv.org/abs/2607.13332}
}

SENTINEL: Stagewise Integrity Verification for Pipeline Parallel Decentralized Training

H. Mohaghegh Dolatabadi, T. Ajanthan, S. Ramasinghe, C. P. Hewa Koneputugodage, G. Avraham, Y. Zuo, V. Shevchenko, A. Long

PDFarXiv
BibTeX
@article{dolatabadi_sentinel_arxiv26,
  title={{SENTINEL: Stagewise Integrity Verification for Pipeline Parallel Decentralized Training}},
  author={Mohaghegh Dolatabadi, Hadi and Ajanthan, Thalaiyasingam and Ramasinghe, Sameera and P Hewa Koneputugodage, Chamin and Avraham, Gil and Zuo, Yan and Shevchenko, Violetta and Long, Alexander},
  year={2026},
  journal={arXiv},
  eprint={2603.03592},
  archiveprefix={arXiv},
  url={https://arxiv.org/abs/2603.03592}
}

2025

NeurIPS

Mixtures of Subspaces for Bandwidth Efficient Context Parallel Training

S. Ramasinghe, T. Ajanthan, H. Mohaghegh Dolatabadi, G. Avraham, V. Shevchenko, Y. Zuo, C. P. Hewa Koneputugodage, A. Long

PDFarXiv
BibTeX
@inproceedings{ramasinghe_mixtures_neurips25,
  title={{Mixtures of Subspaces for Bandwidth Efficient Context Parallel Training}},
  author={Ramasinghe, Sameera and Ajanthan, Thalaiyasingam and Mohaghegh Dolatabadi, Hadi and Avraham, Gil and Shevchenko, Violetta and Zuo, Yan and P Hewa Koneputugodage, Chamin and Long, Alexander},
  year={2025},
  booktitle={NeurIPS}
}

Subspace Networks: Scaling Decentralized Training with Communication-Efficient Model Parallelism

S. Ramasinghe, T. Ajanthan, G. Avraham, Y. Zuo, A. Long

PDFarXiv
BibTeX
@inproceedings{ramasinghe_protocol_neurips25,
  title={{Subspace Networks: Scaling Decentralized Training with Communication-Efficient Model Parallelism}},
  author={Ramasinghe, Sameera and Ajanthan, Thalaiyasingam and Avraham, Gil and Zuo, Yan and Long, Alexander},
  year={2025},
  booktitle={NeurIPS}
}

Unextractable Protocol Models: Collaborative Training and Inference without Weight Materialization

A. Long*, C. P. Hewa Koneputugodage*, T. Ajanthan, Y. Zuo, G. Avraham, V. Shevchenko, H. Mohaghegh Dolatabadi, S. Ramasinghe

PDF
BibTeX
@inproceedings{long_upms_neurips25,
  title={{Unextractable Protocol Models: Collaborative Training and Inference without Weight Materialization}},
  author={Long, Alexander and P Hewa Koneputugodage, Chamin and Ajanthan, Thalaiyasingam and Zuo, Yan and Avraham, Gil and Shevchenko, Violetta and Mohaghegh Dolatabadi, Hadi and Ramasinghe, Sameera},
  year={2025},
  booktitle={NeurIPS}
}

ICML

ICCV

IJCNN

ICLR Workshop

Beyond Top-K: Structured Sparsification for Compression in Pipeline Parallel

S. Ramasinghe, T. Ajanthan, G. Avraham, Y. Zuo, A. Long

PDF
BibTeX
@inproceedings{ramasinghe_beyondtopk_iclr25,
  title={{Beyond Top-K: Structured Sparsification for Compression in Pipeline Parallel}},
  author={Ramasinghe, Sameera and Ajanthan, Thalaiyasingam and Avraham, Gil and Zuo, Yan and Long, Alexander},
  year={2025},
  booktitle={ICLR Workshop: Modularity for Collaborative, Decentralized, and Continual Deep Learning}
}

Exploring Asynchronism in SWARM Parallelism

Y. Zuo, G. Avraham, T. Ajanthan, S. Ramasinghe, A. Long

PDF
BibTeX
@inproceedings{zuo_asyncswarm_iclr25,
  title={{Exploring Asynchronism in SWARM Parallelism}},
  author={Zuo, Yan and Avraham, Gil and Ajanthan, Thalaiyasingam and Ramasinghe, Sameera and Long, Alexander},
  year={2025},
  booktitle={ICLR Workshop: Modularity for Collaborative, Decentralized, and Continual Deep Learning}
}

Momentum Look-Ahead for Asynchronous Distributed Low-Communication Training

T. Ajanthan, S. Ramasinghe, G. Avraham, Y. Zuo, A. Long

PDF
BibTeX
@inproceedings{ajanthan_mla_iclr25,
  title={{Momentum Look-Ahead for Asynchronous Distributed Low-Communication Training}},
  author={Ajanthan, Thalaiyasingam and Ramasinghe, Sameera and Avraham, Gil and Zuo, Yan and Long, Alexander},
  year={2025},
  booktitle={ICLR Workshop: Modularity for Collaborative, Decentralized, and Continual Deep Learning}
}

2024

NeurIPS

CVPR

ICLR Workshop

2023

ICCV

Semi-Supervised Semantic Segmentation under Label Noise via Diverse Learning Groups

P. Li, P. Purkait, T. Ajanthan, M. Abdolshah, R. Garg, H. Husain, C. Xu, S. Gould, W. Ouyang, A. van den Hengel

PDF
BibTeX
@inproceedings{li_dlg_iccv23,
  title={{Semi-Supervised Semantic Segmentation under Label Noise via Diverse Learning Groups}},
  author={Li, Peixia and Purkait, Pulak and Ajanthan, Thalaiyasingam and Abdolshah, Majid and Garg, Ravi and Husain, Hisham and Xu, Chenchen and Gould, Stephen and Ouyang, Wanli and van den Hengel, Anton},
  year={2023},
  booktitle={ICCV}
}

NN

ICLR Workshop

Preprints

Adaptive Cross Batch Normalization for Metric Learning

T. Ajanthan, M. Ma, A. van den Hengel, S. Gould

PDFarXiv
BibTeX
@article{ajanthan_axbn_arxiv23,
  title={{Adaptive Cross Batch Normalization for Metric Learning}},
  author={Ajanthan, Thalaiyasingam and Ma, Matt and van den Hengel, Anton and Gould, Stephen},
  year={2023},
  journal={arXiv},
  eprint={2303.17127},
  archiveprefix={arXiv},
  url={https://arxiv.org/abs/2303.17127}
}

2022

CVPR

Retrieval Augmented Classification for Long-tail Visual Recognition

A. Long, W. Yin, T. Ajanthan, V. Nguyen, P. Purkait, R. Garg, A. Blair, C. Shen, A. van den Hengel

PDFarXiv
BibTeX
@inproceedings{long_rac_cvpr22,
  title={{Retrieval Augmented Classification for Long-tail Visual Recognition}},
  author={Long, Alexander and Yin, Wei and Ajanthan, Thalaiyasingam and Nguyen, Vu and Purkait, Pulak and Garg, Ravi and Blair, Alan and Shen, Chunhua and van den Hengel, Anton},
  year={2022},
  booktitle={CVPR}
}

AAAI

WACV

NeurIPS Workshop

Understanding and Improving the Role of Projection Head in Self-Supervised Learning

K. Gupta, T. Ajanthan, A. van den Hengel, S. Gould

PDF
BibTeX
@inproceedings{gupta_ssl_neurips22,
  title={{Understanding and Improving the Role of Projection Head in Self-Supervised Learning}},
  author={Gupta, Kartik and Ajanthan, Thalaiyasingam and van den Hengel, Anton and Gould, Stephen},
  year={2022},
  booktitle={NeurIPS Workshop: Self-Supervised Learning - Theory and Practice}
}

AAAI Workshop

Provable Defense Against Clustering Attacks on 3D Point Clouds

D. D. Denipitiyage, T. Ajanthan, P. Kamalaruban, A. Weller

PDF
BibTeX
@inproceedings{denipitiyage_pcrobust_aaai22,
  title={{Provable Defense Against Clustering Attacks on 3D Point Clouds}},
  author={Denipitiyage, Dishanika Dewani and Ajanthan, Thalaiyasingam and Kamalaruban, Parameswaran and Weller, Adrian},
  year={2022},
  booktitle={AAAI Workshop: Adversarial Machine Learning and Beyond}
}

2021

ICLR

Calibration of Neural Networks using Splines

K. Gupta, A. Rahimi, T. Ajanthan, T. Mensink, C. Sminchisescu, R. Hartley

PDFarXiv
BibTeX
@inproceedings{gupta_spline_iclr21,
  title={{Calibration of Neural Networks using Splines}},
  author={Gupta, Kartik and Rahimi, Amir and Ajanthan, Thalaiyasingam and Mensink, Thomas and Sminchisescu, Cristian and Hartley, Richard},
  year={2021},
  booktitle={ICLR}
}

AISTATS

DICTA

2020

ICLR

ECCV

3DV

ACCV

IJCV

DGPose: Deep Generative Models for Human Body Analysis

R. De Bem, A. Ghosh, T. Ajanthan, O. Miksik, A. Boukhayma, S. Narayanaswamy, P. H. Torr

PDFarXiv
BibTeX
@article{debem_dgpose_ijcv2020,
  title={{DGPose: Deep Generative Models for Human Body Analysis}},
  author={De Bem, Rodrigo and Ghosh, Arnab and Ajanthan, Thalaiyasingam and Miksik, Ondrej and Boukhayma, Adnane and Narayanaswamy, Siddharth and Torr, Philip HS},
  year={2020},
  journal={IJCV}
}

Preprints

Post-hoc Calibration of Neural Networks

A. Rahimi, K. Gupta, T. Ajanthan, T. Mensink, C. Sminchisescu, R. Hartley

PDFarXiv
BibTeX
@article{rahimi_glayers_arxiv20,
  title={{Post-hoc Calibration of Neural Networks}},
  author={Rahimi, Amir and Gupta, Kartik and Ajanthan, Thalaiyasingam and Mensink, Thomas and Sminchisescu, Cristian and Hartley, Richard},
  year={2020},
  journal={arXiv},
  eprint={2006.12807},
  archiveprefix={arXiv},
  url={https://arxiv.org/abs/2006.12807}
}

2019

ICLR

CVPR

Learning to Adapt for Stereo

A. Tonioni, O. Rahnama*, T. Joy*, L. Di Stefano, T. Ajanthan, P. H. Torr

PDFarXivCode
BibTeX
@inproceedings{tonioni_l2a_cvpr19,
  title={{Learning to Adapt for Stereo}},
  author={Tonioni, Alessio and Rahnama, Oscar and Joy, Thomas and Di Stefano, Luigi and Ajanthan, Thalaiyasingam and Torr, Philip HS},
  year={2019},
  booktitle={CVPR}
}

ICCV

WACV

A Conditional Deep Generative Model of People in Natural Images

R. De Bem, A. Ghosh, A. Boukhayma, T. Ajanthan, S. Narayanaswamy, P. H. Torr

PDF
BibTeX
@inproceedings{debem_dgpose_wacv2019,
  title={{A Conditional Deep Generative Model of People in Natural Images}},
  author={De Bem, Rodrigo and Ghosh, Arnab and Boukhayma, Adnane and Ajanthan, Thalaiyasingam and Narayanaswamy, Siddharth and Torr, Philip HS},
  year={2019},
  booktitle={WACV}
}

SIIMS

Efficient Relaxations for Dense CRFs with Sparse Higher Order Potentials

T. Joy*, A. Desmaison*, T. Ajanthan*, R. Bunel, M. Salzmann, P. Kohli, P. H. Torr, M. P. Kumar

PDFarXivCode
BibTeX
@article{joy_higherorder_siam2018,
  title={{Efficient Relaxations for Dense CRFs with Sparse Higher Order Potentials}},
  author={Joy, Thomas and Desmaison, Alban and Ajanthan, Thalaiyasingam and Bunel, Rudy and Salzmann, Mathieu and Kohli, Pushmeet and Torr, Philip HS and Kumar, M Pawan},
  year={2019},
  journal={SIIMS}
}

ICML Workshop

1500+ citations

Continual Learning with Tiny Episodic Memories

A. Chaudhry, M. Rohrbach, M. Elhoseiny, T. Ajanthan, P. K. Dokania, P. H. Torr, M. A. Ranzato

PDFarXiv
BibTeX
@inproceedings{chaudhry_tinyer_icml2019,
  title={{Continual Learning with Tiny Episodic Memories}},
  author={Chaudhry, Arslan and Rohrbach, Marcus and Elhoseiny, Mohamed and Ajanthan, Thalaiyasingam and Dokania, Puneet K and Torr, Philip HS and Ranzato, Marc'Aurelio},
  year={2019},
  booktitle={ICML Workshop: Multi-Task and Lifelong Reinforcement Learning}
}

2018

ECCV

PAMI

CVPR Workshop

ECCV Workshop

oral

A Semi-supervised Deep Generative Model for Human Body Analysis

R. De Bem, A. Ghosh, T. Ajanthan, O. Miksik, S. Narayanaswamy, P. H. Torr

PDF
BibTeX
@inproceedings{debem_dgpose_eccv2018,
  title={{A Semi-supervised Deep Generative Model for Human Body Analysis}},
  author={De Bem, Rodrigo and Ghosh, Arnab and Ajanthan, Thalaiyasingam and Miksik, Ondrej and Narayanaswamy, Siddharth and Torr, Philip HS},
  year={2018},
  booktitle={ECCV Workshop on Human Behaviour Understanding}
}

Preprints

Generalized Range Moves

R. Hartley, T. Ajanthan

PDFarXiv
BibTeX
@article{hartley_gswap_arxiv18,
  title={{Generalized Range Moves}},
  author={Hartley, Richard and Ajanthan, Thalaiyasingam},
  year={2018},
  journal={arXiv},
  eprint={1811.09171},
  archiveprefix={arXiv},
  url={https://arxiv.org/abs/1811.09171}
}

2017

CVPR

Efficient Linear Programming for Dense CRFs

T. Ajanthan, A. Desmaison, R. Bunel, M. Salzmann, P. H. Torr, M. P. Kumar

PDFarXivCodeSlides
BibTeX
@inproceedings{ajanthan_proxlp_cvpr2017,
  title={{Efficient Linear Programming for Dense CRFs}},
  author={Ajanthan, Thalaiyasingam and Desmaison, Alban and Bunel, Rudy and Salzmann, Mathieu and Torr, Philip HS and Kumar, M Pawan},
  year={2017},
  booktitle={CVPR}
}

Thesis

2016

CVPR

2015

CVPR

2013

ICIIS

* Equal contribution