% Linear and sparse attention bibliography
% Curated and verified on 2026-07-21.
% Total: 73 entries.
%   - 20 linear-attention and adjacent references from the previously curated file.
%   - 18 normalized records from the supplied sparse-attention file.
%   - 35 added sparse-attention references selected for historical, algorithmic,
%     kernel, hardware, serving-system, long-context LLM, and video-generation coverage.
%
% Bibliographic policy:
%   - Prefer the formally published conference/journal version when verified.
%   - Retain arXiv identifiers for discoverability and use arXiv as the primary
%     record when no verified proceedings version was available.
%   - FlashAttention 1--4 are exact-attention baselines, retained because sparse
%     attention performance is normally evaluated against them.
%   - Miccini et al. (2024) is a GRU quantization paper, retained as an adjacent
%     methods reference inherited from the linear-attention bibliography.
%
% Organization:
%   Part I:  Linear attention (sections 1--4)
%   Part II: Sparse attention and exact-attention baselines (sections 5--10)

% ============================================================================
% PART I. LINEAR ATTENTION
% ============================================================================
% ============================================================================
% 1. Foundations and early formulations
% ============================================================================

@inproceedings{katharopoulos2020transformers,
  title     = {Transformers are {RNNs}: Fast Autoregressive Transformers with Linear Attention},
  author    = {Katharopoulos, Angelos and Vyas, Apoorv and Pappas, Nikolaos and Fleuret, Fran{\c{c}}ois},
  booktitle = {Proceedings of the 37th International Conference on Machine Learning},
  series    = {Proceedings of Machine Learning Research},
  volume    = {119},
  pages     = {5156--5165},
  publisher = {PMLR},
  year      = {2020},
  url       = {https://proceedings.mlr.press/v119/katharopoulos20a.html}
}

@inproceedings{choromanski2021rethinking,
  title         = {Rethinking Attention with Performers},
  author        = {Choromanski, Krzysztof and Likhosherstov, Valerii and Dohan, David and Song, Xingyou and Gane, Andreea and Sarlos, Tamas and Hawkins, Peter and Davis, Jared and Mohiuddin, Afroz and Kaiser, Lukasz and Belanger, David and Colwell, Lucy and Weller, Adrian},
  booktitle     = {International Conference on Learning Representations},
  year          = {2021},
  eprint        = {2009.14794},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2009.14794},
  url           = {https://openreview.net/forum?id=Ua6zuk0WRH}
}

@inproceedings{schlag2021linear,
  title     = {Linear Transformers Are Secretly Fast Weight Programmers},
  author    = {Schlag, Imanol and Irie, Kazuki and Schmidhuber, J{\"u}rgen},
  booktitle = {Proceedings of the 38th International Conference on Machine Learning},
  series    = {Proceedings of Machine Learning Research},
  volume    = {139},
  pages     = {9355--9366},
  publisher = {PMLR},
  year      = {2021},
  url       = {https://proceedings.mlr.press/v139/schlag21a.html}
}

@inproceedings{hua2022transformer,
  title     = {Transformer Quality in Linear Time},
  author    = {Hua, Weizhe and Dai, Zihang and Liu, Hanxiao and Le, Quoc},
  booktitle = {Proceedings of the 39th International Conference on Machine Learning},
  series    = {Proceedings of Machine Learning Research},
  volume    = {162},
  pages     = {9099--9117},
  publisher = {PMLR},
  year      = {2022},
  url       = {https://proceedings.mlr.press/v162/hua22a.html}
}

@article{sun2023retentive,
  title         = {Retentive Network: A Successor to Transformer for Large Language Models},
  author        = {Sun, Yutao and Dong, Li and Huang, Shaohan and Ma, Shuming and Xia, Yuqing and Xue, Jilong and Wang, Jianyong and Wei, Furu},
  journal       = {arXiv preprint arXiv:2307.08621},
  year          = {2023},
  eprint        = {2307.08621},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2307.08621},
  url           = {https://arxiv.org/abs/2307.08621}
}

% ============================================================================
% 2. Modern gated, delta-rule, SSD, and hybrid architectures
% ============================================================================

@inproceedings{yang2024gated,
  title     = {Gated Linear Attention Transformers with Hardware-Efficient Training},
  author    = {Yang, Songlin and Wang, Bailin and Shen, Yikang and Panda, Rameswar and Kim, Yoon},
  booktitle = {Proceedings of the 41st International Conference on Machine Learning},
  series    = {Proceedings of Machine Learning Research},
  volume    = {235},
  pages     = {56501--56523},
  publisher = {PMLR},
  year      = {2024},
  eprint    = {2312.06635},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi       = {10.48550/arXiv.2312.06635},
  url       = {https://proceedings.mlr.press/v235/yang24ab.html}
}

@inproceedings{yang2024delta,
  title         = {Parallelizing Linear Transformers with the Delta Rule over Sequence Length},
  author        = {Yang, Songlin and Wang, Bailin and Zhang, Yu and Shen, Yikang and Kim, Yoon},
  booktitle     = {Advances in Neural Information Processing Systems},
  volume        = {37},
  year          = {2024},
  eprint        = {2406.06484},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2406.06484},
  url           = {https://arxiv.org/abs/2406.06484}
}

@inproceedings{dao2024transformers,
  title     = {Transformers are {SSMs}: Generalized Models and Efficient Algorithms Through Structured State Space Duality},
  author    = {Dao, Tri and Gu, Albert},
  booktitle = {Proceedings of the 41st International Conference on Machine Learning},
  series    = {Proceedings of Machine Learning Research},
  volume    = {235},
  pages     = {10041--10071},
  publisher = {PMLR},
  year      = {2024},
  url       = {https://proceedings.mlr.press/v235/dao24a.html}
}

@inproceedings{arora2024simple,
  title     = {Simple Linear Attention Language Models Balance the Recall--Throughput Tradeoff},
  author    = {Arora, Simran and Eyuboglu, Sabri and Zhang, Michael and Timalsina, Aman and Alberti, Silas and Zou, James and Rudra, Atri and R{\'e}, Christopher},
  booktitle = {Proceedings of the 41st International Conference on Machine Learning},
  series    = {Proceedings of Machine Learning Research},
  volume    = {235},
  pages     = {1763--1840},
  publisher = {PMLR},
  year      = {2024},
  url       = {https://proceedings.mlr.press/v235/arora24a.html}
}

@inproceedings{yang2025gated,
  title         = {Gated Delta Networks: Improving {Mamba2} with Delta Rule},
  author        = {Yang, Songlin and Kautz, Jan and Hatamizadeh, Ali},
  booktitle     = {The Thirteenth International Conference on Learning Representations},
  year          = {2025},
  eprint        = {2412.06464},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2412.06464},
  url           = {https://openreview.net/forum?id=r8H7xhYPwz}
}

@article{kimi2025linear,
  title  = {{Kimi Linear}: An Expressive, Efficient Attention Architecture},
  author = {Zhang, Yu and Lin, Zongyu and Yao, Xingcheng and Hu, Jiaxi and Meng, Fanqing and Liu, Chengyin and Men, Xin and Yang, Songlin and Li, Zhiyuan and Li, Wentao and Lu, Enzhe and Liu, Weizhou and Chen, Yanru and Xu, Weixin and Yu, Longhui and Wang, Yejie and Fan, Yu and Zhong, Longguang and Yuan, Enming and Zhang, Dehao and Zhang, Yizhi and Liu, T. Y. and Wang, Haiming and Fang, Shengjun and He, Weiran and Liu, Shaowei and Li, Yiwei and Su, Jianlin and Qiu, Jiezhong and Pang, Bo and Yan, Junjie and Jiang, Zhejun and Huang, Weixiao and Yin, Bohong and You, Jiacheng and Wei, Chu and Wang, Zhengtao and Hong, Chao and Chen, Yutian and Chen, Guanduo and Wang, Yucheng and Zheng, Huabin and Wang, Feng and Liu, Yibo and Dong, Mengnan and Zhang, Zheng and Pan, Siyuan and Wu, Wenhao and Wu, Yuhao and Guan, Longyu and Tao, Jiawen and Fu, Guohong and Xu, Xinran and Wang, Yuzhi and Lai, Guokun and Wu, Yuxin and Zhou, Xinyu and Yang, Zhilin and Du, Yulun},
  journal       = {arXiv preprint arXiv:2510.26692},
  year          = {2025},
  eprint        = {2510.26692},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2510.26692},
  url           = {https://arxiv.org/abs/2510.26692},
  note          = {Kimi Team technical report}
}

@article{huang2026mdn,
  title         = {{MDN}: Parallelizing Stepwise Momentum for Delta Linear Attention},
  author        = {Huang, Yulong and Liu, Xiang and Huang, Hongxiang and Lin, Xiaopeng and Liu, Zunchang and Chu, Xiaowen and Xie, Zeke and Cheng, Bojun},
  journal       = {arXiv preprint arXiv:2605.05838},
  year          = {2026},
  eprint        = {2605.05838},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2605.05838},
  url           = {https://arxiv.org/abs/2605.05838}
}

% ============================================================================
% 3. GPU kernels and distributed systems
% ============================================================================

@inproceedings{qin2024various,
  title     = {Various Lengths, Constant Speed: Efficient Language Modeling with Lightning Attention},
  author    = {Qin, Zhen and Sun, Weigao and Li, Dong and Shen, Xuyang and Sun, Weixuan and Zhong, Yiran},
  booktitle = {Proceedings of the 41st International Conference on Machine Learning},
  series    = {Proceedings of Machine Learning Research},
  volume    = {235},
  pages     = {41517--41535},
  publisher = {PMLR},
  year      = {2024},
  url       = {https://proceedings.mlr.press/v235/qin24c.html}
}

@article{sun2025lasp,
  title         = {{LASP}: Linear Attention Sequence Parallelism},
  author        = {Sun, Weigao and Qin, Zhen and Li, Dong and Shen, Xuyang and Qiao, Yu and Zhong, Yiran},
  journal       = {Transactions on Machine Learning Research},
  year          = {2025},
  eprint        = {2404.02882},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2404.02882},
  url           = {https://openreview.net/forum?id=gG8sQUUtN7}
}

@article{sun2025lasp2,
  title         = {{LASP-2}: Rethinking Sequence Parallelism for Linear Attention and Its Hybrid},
  author        = {Sun, Weigao and Lan, Disen and Zhong, Yiran and Qu, Xiaoye and Cheng, Yu},
  journal       = {arXiv preprint arXiv:2502.07563},
  year          = {2025},
  eprint        = {2502.07563},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2502.07563},
  url           = {https://arxiv.org/abs/2502.07563},
  note          = {Technical report}
}

@inproceedings{beck2025tiled,
  title         = {Tiled Flash Linear Attention: More Efficient Linear {RNN} and {xLSTM} Kernels},
  author        = {Beck, Maximilian and P{\"o}ppel, Korbinian and Lippe, Phillip and Hochreiter, Sepp},
  booktitle     = {Advances in Neural Information Processing Systems},
  volume        = {38},
  year          = {2025},
  eprint        = {2503.14376},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2503.14376},
  url           = {https://proceedings.neurips.cc/paper_files/paper/2025/hash/6cb81234ab47027e991728ed7dd76735-Abstract-Conference.html}
}

@article{gerami2025transformer,
  title         = {Transformer Based Linear Attention with Optimized {GPU} Kernel Implementation},
  author        = {Gerami, Armin and Duraiswami, Ramani},
  journal       = {arXiv preprint arXiv:2510.21956},
  year          = {2025},
  eprint        = {2510.21956},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2510.21956},
  url           = {https://arxiv.org/abs/2510.21956}
}

% ============================================================================
% 4. Quantization and state reduction
% ============================================================================

@inproceedings{miccini2024towards,
  title         = {Towards a Tailored Mixed-Precision Sub-8-Bit Quantization Scheme for Gated Recurrent Units Using Genetic Algorithms},
  author        = {Miccini, Riccardo and Cerioli, Alessandro and Laroche, Cl{\'e}ment and Piechowiak, Tobias and Spars{\o}, Jens and Pezzarossa, Luca},
  booktitle     = {tinyML Research Symposium 2024},
  year          = {2024},
  eprint        = {2402.12263},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2402.12263},
  url           = {https://arxiv.org/abs/2402.12263}
}

@inproceedings{kim2026ssdi8,
  title     = {{SSDi8}: Accurate and Efficient 8-bit Quantization for State Space Duality},
  author    = {Kim, Hyunwoo and Ko, Byoungchan and Kang, Minseok and Kim, Minwoo and Lee, Dongjin and Lee, Jaehoon and Yoon, Sungroh and Jung, Dahuin},
  booktitle = {The Fourteenth International Conference on Learning Representations},
  year      = {2026},
  url       = {https://openreview.net/forum?id=pjMDZJd4rT}
}

@article{nazari2026key,
  title         = {The Key to State Reduction in Linear Attention: A Rank-Based Perspective},
  author        = {Nazari, Philipp and Rusch, T. Konstantin},
  journal       = {arXiv preprint arXiv:2602.04852},
  year          = {2026},
  eprint        = {2602.04852},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2602.04852},
  url           = {https://arxiv.org/abs/2602.04852}
}

% ============================================================================
% PART II. SPARSE ATTENTION AND EXACT-ATTENTION BASELINES
% ============================================================================

% ============================================================================
% 5. Foundations and classic sparse-attention formulations
% ============================================================================

@article{child2019sparse,
  title         = {Generating Long Sequences with Sparse Transformers},
  author        = {Child, Rewon and Gray, Scott and Radford, Alec and Sutskever, Ilya},
  journal       = {arXiv preprint arXiv:1904.10509},
  year          = {2019},
  eprint        = {1904.10509},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.1904.10509},
  url           = {https://arxiv.org/abs/1904.10509}
}

@inproceedings{kitaev2020reformer,
  title         = {Reformer: The Efficient Transformer},
  author        = {Kitaev, Nikita and Kaiser, {\L}ukasz and Levskaya, Anselm},
  booktitle     = {International Conference on Learning Representations},
  year          = {2020},
  eprint        = {2001.04451},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2001.04451},
  url           = {https://arxiv.org/abs/2001.04451}
}

@article{beltagy2020longformer,
  title         = {Longformer: The Long-Document Transformer},
  author        = {Beltagy, Iz and Peters, Matthew E. and Cohan, Arman},
  journal       = {arXiv preprint arXiv:2004.05150},
  year          = {2020},
  eprint        = {2004.05150},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2004.05150},
  url           = {https://arxiv.org/abs/2004.05150}
}

@inproceedings{zaheer2020bigbird,
  title         = {{Big Bird}: Transformers for Longer Sequences},
  author        = {Zaheer, Manzil and Guruganesh, Guru and Dubey, Kumar Avinava and Ainslie, Joshua and Alberti, Chris and Onta{\~n}{\'o}n, Santiago and Pham, Philip and Ravula, Anirudh and Wang, Qifan and Yang, Li and Ahmed, Amr},
  booktitle     = {Advances in Neural Information Processing Systems},
  volume        = {33},
  pages         = {17283--17297},
  year          = {2020},
  eprint        = {2007.14062},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2007.14062},
  url           = {https://arxiv.org/abs/2007.14062}
}

@inproceedings{ainslie2020etc,
  title     = {ETC: Encoding Long and Structured Inputs in Transformers},
  author    = {Ainslie, Joshua and Onta{\~n}{\'o}n, Santiago and Alberti, Chris and Cvicek, Vaclav and Fisher, Zachary and Pham, Philip and Ravula, Anirudh and Sanghai, Sumit and Wang, Qifan and Yang, Li},
  booktitle = {Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing},
  pages     = {268--284},
  publisher = {Association for Computational Linguistics},
  year      = {2020},
  doi       = {10.18653/v1/2020.emnlp-main.19},
  url       = {https://aclanthology.org/2020.emnlp-main.19/}
}

@inproceedings{tay2020sparse,
  title     = {Sparse Sinkhorn Attention},
  author    = {Tay, Yi and Bahri, Dara and Yang, Liu and Metzler, Donald and Juan, Da-Cheng},
  booktitle = {Proceedings of the 37th International Conference on Machine Learning},
  series    = {Proceedings of Machine Learning Research},
  volume    = {119},
  pages     = {9438--9447},
  publisher = {PMLR},
  year      = {2020},
  url       = {https://proceedings.mlr.press/v119/tay20a.html}
}

@article{roy2021routing,
  title   = {Efficient Content-Based Sparse Attention with Routing Transformers},
  author  = {Roy, Aurko and Saffar, Mohammad and Vaswani, Ashish and Grangier, David},
  journal = {Transactions of the Association for Computational Linguistics},
  volume  = {9},
  pages   = {53--68},
  year    = {2021},
  doi     = {10.1162/tacl_a_00353},
  url     = {https://aclanthology.org/2021.tacl-1.4/}
}

@inproceedings{chen2021scatterbrain,
  title     = {{Scatterbrain}: Unifying Sparse and Low-Rank Attention Approximation},
  author    = {Chen, Beidi and Dao, Tri and Winsor, Eric and Song, Zhao and Rudra, Atri and R{\'e}, Christopher},
  booktitle = {Advances in Neural Information Processing Systems},
  volume    = {34},
  pages     = {17413--17426},
  year      = {2021},
  eprint        = {2110.15343},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2110.15343},
  url           = {https://proceedings.neurips.cc/paper/2021/hash/9185f3ec501c674c7c788464a36e7fb3-Abstract.html}
}

@article{ding2023longnet,
  title         = {{LongNet}: Scaling Transformers to 1,000,000,000 Tokens},
  author        = {Ding, Jiayu and Ma, Shuming and Dong, Li and Zhang, Xingxing and Huang, Shaohan and Wang, Wenhui and Zheng, Nanning and Wei, Furu},
  journal       = {arXiv preprint arXiv:2307.02486},
  year          = {2023},
  eprint        = {2307.02486},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2307.02486},
  url           = {https://arxiv.org/abs/2307.02486}
}

@inproceedings{han2024hyperattention,
  title         = {{HyperAttention}: Long-Context Attention in Near-Linear Time},
  author        = {Han, Insu and Jayaram, Rajesh and Karbasi, Amin and Mirrokni, Vahab and Woodruff, David P. and Zandieh, Amir},
  booktitle     = {The Twelfth International Conference on Learning Representations},
  year          = {2024},
  eprint        = {2310.05869},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2310.05869},
  url           = {https://arxiv.org/abs/2310.05869}
}

% ============================================================================
% 6. Hardware, kernels, and serving systems
% ============================================================================

@inproceedings{wang2021spatten,
  title     = {{SpAtten}: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning},
  author    = {Wang, Hanrui and Zhang, Zhekai and Han, Song},
  booktitle = {2021 IEEE International Symposium on High-Performance Computer Architecture (HPCA)},
  pages     = {97--110},
  publisher = {IEEE},
  year      = {2021},
  doi       = {10.1109/HPCA51647.2021.00018},
  url       = {https://ieeexplore.ieee.org/document/9407232/}
}

@inproceedings{lu2021sanger,
  title     = {{Sanger}: A Co-Design Framework for Enabling Sparse Attention Using Reconfigurable Architecture},
  author    = {Lu, Liqiang and Jin, Yicheng and Bi, Hangrui and Luo, Zizhang and Li, Peng and Wang, Tao and Liang, Yun},
  booktitle = {2021 54th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO)},
  pages     = {977--991},
  publisher = {IEEE},
  year      = {2021},
  doi       = {10.1145/3466752.3480125},
  url       = {https://doi.org/10.1145/3466752.3480125}
}

@inproceedings{shen2022salo,
  title     = {{SALO}: An Efficient Spatial Accelerator Enabling Hybrid Sparse Attention Mechanisms for Long Sequences},
  author    = {Shen, Guan and Zhao, Jieru and Chen, Quan and Leng, Jingwen and Li, Chao and Guo, Minyi},
  booktitle = {2022 59th ACM/IEEE Design Automation Conference (DAC)},
  pages     = {571--576},
  publisher = {ACM},
  year      = {2022},
  doi       = {10.1145/3489517.3530504},
  url       = {https://doi.org/10.1145/3489517.3530504}
}

@inproceedings{pagliardini2023dynamic,
  title         = {Fast Attention over Long Sequences with Dynamic Sparse Flash Attention},
  author        = {Pagliardini, Matteo and Paliotta, Daniele and Jaggi, Martin and Fleuret, Fran{\c{c}}ois},
  booktitle     = {Advances in Neural Information Processing Systems},
  volume        = {36},
  pages         = {59808--59831},
  year          = {2023},
  eprint        = {2306.01160},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2306.01160},
  url           = {https://proceedings.neurips.cc/paper_files/paper/2023/hash/bc222e8153a49c1b30a1b8ba96b35117-Abstract-Conference.html}
}

@inproceedings{dong2025flexattention,
  title         = {{FlexAttention}: A Programming Model for Generating Fused Attention Variants},
  author        = {Dong, Juechu and Feng, Boyuan and Guessous, Driss and Liang, Yanbo and He, Horace},
  booktitle     = {Proceedings of Machine Learning and Systems},
  volume        = {7},
  year          = {2025},
  eprint        = {2412.05496},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2412.05496},
  url           = {https://proceedings.mlsys.org/paper_files/paper/2025/hash/61a9278dfef5f871b5e472389f8d6fa1-Abstract-Conference.html}
}

@inproceedings{ye2025flashinfer,
  title         = {{FlashInfer}: Efficient and Customizable Attention Engine for {LLM} Inference Serving},
  author        = {Ye, Zihao and Chen, Lequn and Lai, Ruihang and Lin, Wuwei and Zhang, Yineng and Wang, Stephanie and Chen, Tianqi and Kasikci, Baris and Grover, Vinod and Krishnamurthy, Arvind and Ceze, Luis},
  booktitle     = {Proceedings of Machine Learning and Systems},
  volume        = {7},
  year          = {2025},
  eprint        = {2501.01005},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2501.01005},
  url           = {https://proceedings.mlsys.org/paper_files/paper/2025/hash/dbf02b21d77409a2db30e56866a8ab3a-Abstract-Conference.html}
}


@inproceedings{lee2024infinigen,
  title         = {{InfiniGen}: Efficient Generative Inference of Large Language Models with Dynamic {KV} Cache Management},
  author        = {Lee, Wonbeom and Lee, Jungi and Seo, Junghwan and Sim, Jaewoong},
  booktitle     = {18th USENIX Symposium on Operating Systems Design and Implementation (OSDI 24)},
  pages         = {155--172},
  publisher     = {USENIX Association},
  address       = {Santa Clara, CA},
  isbn          = {978-1-939133-40-3},
  year          = {2024},
  month         = jul,
  eprint        = {2406.19707},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2406.19707},
  url           = {https://www.usenix.org/conference/osdi24/presentation/lee}
}

@inproceedings{yang2025lserve,
  title         = {{LServe}: Efficient Long-Sequence {LLM} Serving with Unified Sparse Attention},
  author        = {Yang, Shang and Guo, Junxian and Tang, Haotian and Hu, Qinghao and Xiao, Guangxuan and Tang, Jiaming and Lin, Yujun and Liu, Zhijian and Lu, Yao and Han, Song},
  booktitle     = {Proceedings of Machine Learning and Systems},
  volume        = {7},
  year          = {2025},
  eprint        = {2502.14866},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2502.14866},
  url           = {https://arxiv.org/abs/2502.14866}
}

% ============================================================================
% 7. Long-context LLM inference and trainable sparse architectures
% ============================================================================

@inproceedings{xiao2024streamingllm,
  title         = {Efficient Streaming Language Models with Attention Sinks},
  author        = {Xiao, Guangxuan and Tian, Yuandong and Chen, Beidi and Han, Song and Lewis, Mike},
  booktitle     = {The Twelfth International Conference on Learning Representations},
  year          = {2024},
  eprint        = {2309.17453},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2309.17453},
  url           = {https://arxiv.org/abs/2309.17453}
}

@inproceedings{zhang2023h2o,
  title         = {H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models},
  author        = {Zhang, Zhenyu and Sheng, Ying and Zhou, Tianyi and Chen, Tianlong and Zheng, Lianmin and Cai, Ruisi and Song, Zhao and Tian, Yuandong and R{\'e}, Christopher and Barrett, Clark and Wang, Zhangyang and Chen, Beidi},
  booktitle     = {Advances in Neural Information Processing Systems},
  volume        = {36},
  pages         = {34661--34710},
  year          = {2023},
  eprint        = {2306.14048},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2306.14048},
  url           = {https://arxiv.org/abs/2306.14048}
}

@inproceedings{jiang2024minference,
  title         = {{MInference} 1.0: Accelerating Pre-Filling for Long-Context {LLMs} via Dynamic Sparse Attention},
  author        = {Jiang, Huiqiang and Li, Yucheng and Zhang, Chengruidong and Wu, Qianhui and Luo, Xufang and Ahn, Surin and Han, Zhenhua and Abdi, Amir H. and Li, Dongsheng and Lin, Chin-Yew and Yang, Yuqing and Qiu, Lili},
  booktitle     = {Advances in Neural Information Processing Systems},
  volume        = {37},
  pages         = {52481--52515},
  year          = {2024},
  eprint        = {2407.02490},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2407.02490},
  url           = {https://proceedings.neurips.cc/paper_files/paper/2024/hash/5dfbe6f5671e82c76841ba687a8a9ecb-Abstract-Conference.html}
}

@inproceedings{tang2024quest,
  title     = {{QUEST}: Query-Aware Sparsity for Efficient Long-Context {LLM} Inference},
  author    = {Tang, Jiaming and Zhao, Yilong and Zhu, Kan and Xiao, Guangxuan and Kasikci, Baris and Han, Song},
  booktitle = {Proceedings of the 41st International Conference on Machine Learning},
  series    = {Proceedings of Machine Learning Research},
  volume    = {235},
  pages     = {47901--47911},
  publisher = {PMLR},
  year      = {2024},
  url       = {https://proceedings.mlr.press/v235/tang24l.html}
}

@inproceedings{ribar2024sparq,
  title         = {{SparQ} Attention: Bandwidth-Efficient {LLM} Inference},
  author        = {Ribar, Luka and Chelombiev, Ivan and Hudlass-Galley, Luke and Blake, Charlie and Luschi, Carlo and Orr, Douglas},
  booktitle     = {Proceedings of the 41st International Conference on Machine Learning},
  series        = {Proceedings of Machine Learning Research},
  volume        = {235},
  pages         = {42558--42583},
  publisher     = {PMLR},
  year          = {2024},
  eprint        = {2312.04985},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2312.04985},
  url           = {https://proceedings.mlr.press/v235/ribar24a.html}
}

@inproceedings{singhania2024loki,
  title         = {{Loki}: Low-Rank Keys for Efficient Sparse Attention},
  author        = {Singhania, Prajwal and Singh, Siddharth and He, Shwai and Feizi, Soheil and Bhatele, Abhinav},
  booktitle     = {Advances in Neural Information Processing Systems},
  volume        = {37},
  year          = {2024},
  eprint        = {2406.02542},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2406.02542},
  url           = {https://arxiv.org/abs/2406.02542}
}


@inproceedings{chen2025magicpig,
  title         = {{MagicPIG}: {LSH} Sampling for Efficient {LLM} Generation},
  author        = {Chen, Zhuoming and Sadhukhan, Ranajoy and Ye, Zihao and Zhou, Yang and Zhang, Jianyu and Nolte, Niklas and Tian, Yuandong and Douze, Matthijs and Bottou, L{\'e}on and Jia, Zhihao and Chen, Beidi},
  booktitle     = {The Thirteenth International Conference on Learning Representations},
  year          = {2025},
  eprint        = {2410.16179},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2410.16179},
  url           = {https://openreview.net/forum?id=ALzTQUgW8a}
}

@article{liu2024retrievalattention,
  title         = {{RetrievalAttention}: Accelerating Long-Context {LLM} Inference via Vector Retrieval},
  author        = {Liu, Di and Chen, Meng and Lu, Baotong and Jiang, Huiqiang and Han, Zhenhua and Zhang, Qianxi and Chen, Qi and Zhang, Chengruidong and Ding, Bailu and Zhang, Kai and Chen, Chen and Yang, Fan and Yang, Yuqing and Qiu, Lili},
  journal       = {arXiv preprint arXiv:2409.10516},
  year          = {2024},
  eprint        = {2409.10516},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2409.10516},
  url           = {https://arxiv.org/abs/2409.10516}
}

@article{zhu2024sampleattention,
  title         = {{SampleAttention}: Near-Lossless Acceleration of Long-Context {LLM} Inference with Adaptive Structured Sparse Attention},
  author        = {Zhu, Qianchao and Duan, Jiangfei and Chen, Chang and Liu, Siran and Li, Xiuhong and Feng, Guanyu and Lv, Xin and Cao, Huanqi and Xiao, Chuanfu and Zhang, Xingcheng and Lin, Dahua and Yang, Chao},
  journal       = {arXiv preprint arXiv:2406.15486},
  year          = {2024},
  eprint        = {2406.15486},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2406.15486},
  url           = {https://arxiv.org/abs/2406.15486}
}

@article{lou2024sparsek,
  title         = {Sparser Is Faster and Less Is More: Efficient Sparse Attention for Long-Range Transformers},
  author        = {Lou, Chao and Jia, Zixia and Zheng, Zilong and Tu, Kewei},
  journal       = {arXiv preprint arXiv:2406.16747},
  year          = {2024},
  eprint        = {2406.16747},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2406.16747},
  url           = {https://arxiv.org/abs/2406.16747}
}

@inproceedings{lee2025hip,
  title         = {A Training-Free Sub-Quadratic Cost Transformer Model Serving Framework with Hierarchically Pruned Attention},
  author        = {Lee, Heejun and Park, Geon and Lee, Youngwan and Suh, Jaduk and Kim, Jina and Jeong, Wonyoung and Kim, Bumsik and Lee, Hyemin and Jeon, Myeongjae and Hwang, Sung Ju},
  booktitle     = {The Thirteenth International Conference on Learning Representations},
  year          = {2025},
  eprint        = {2406.09827},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2406.09827},
  url           = {https://openreview.net/forum?id=PTcMzQgKmn}
}

@inproceedings{xiao2025duoattention,
  title         = {{DuoAttention}: Efficient Long-Context {LLM} Inference with Retrieval and Streaming Heads},
  author        = {Xiao, Guangxuan and Tang, Jiaming and Zuo, Jingwei and Guo, Junxian and Yang, Shang and Tang, Haotian and Fu, Yao and Han, Song},
  booktitle     = {The Thirteenth International Conference on Learning Representations},
  year          = {2025},
  eprint        = {2410.10819},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2410.10819},
  url           = {https://openreview.net/forum?id=cFu7ze7xUm}
}

@inproceedings{lai2025flexprefill,
  title         = {{FlexPrefill}: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference},
  author        = {Lai, Xunhao and Lu, Jianqiao and Luo, Yao and Ma, Yiyuan and Zhou, Xun},
  booktitle     = {The Thirteenth International Conference on Learning Representations},
  year          = {2025},
  eprint        = {2502.20766},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2502.20766},
  url           = {https://openreview.net/forum?id=r5GJDVJHmr}
}

@article{yuan2025native,
  title         = {Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention},
  author        = {Yuan, Jingyang and Gao, Huazuo and Dai, Damai and Luo, Junyu and Zhao, Liang and Zhang, Zhengyan and Xie, Zhenda and Wei, Y. X. and Wang, Lean and Xiao, Zhiping and Wang, Yuqing and Ruan, Chong and Zhang, Ming and Liang, Wenfeng and Zeng, Wangding},
  journal       = {arXiv preprint arXiv:2502.11089},
  year          = {2025},
  eprint        = {2502.11089},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2502.11089},
  url           = {https://arxiv.org/abs/2502.11089}
}

@inproceedings{gao2025seerattention,
  title         = {{SeerAttention}: Self-Distilled Attention Gating for Efficient Long-Context Prefilling},
  author        = {Gao, Yizhao and Zeng, Zhichen and Du, Dayou and Cao, Shijie and Zhou, Peiyuan and Qi, Jiaxing and Lai, Junjie and So, Hayden Kwok-Hay and Cao, Ting and Yang, Fan and Yang, Mao},
  booktitle     = {Advances in Neural Information Processing Systems},
  volume        = {38},
  year          = {2025},
  eprint        = {2410.13276},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2410.13276},
  url           = {https://proceedings.neurips.cc/paper_files/paper/2025/hash/50e9dbc4ab68d94f15261ddc26c8ca2b-Abstract-Conference.html}
}

@article{lu2025moba,
  title         = {{MoBA}: Mixture of Block Attention for Long-Context {LLMs}},
  author        = {Lu, Enzhe and Jiang, Zhejun and Liu, Jingyuan and Du, Yulun and Jiang, Tao and Hong, Chao and Liu, Shaowei and He, Weiran and Yuan, Enming and Wang, Yuzhi and Huang, Zhiqi and Yuan, Huan and Xu, Suting and Xu, Xinran and Lai, Guokun and Chen, Yanru and Zheng, Huabin and Yan, Junjie and Su, Jianlin and Wu, Yuxin and Zhang, Neo Y. and Yang, Zhilin and Zhou, Xinyu and Zhang, Mingxing and Qiu, Jiezhong},
  journal       = {arXiv preprint arXiv:2502.13189},
  year          = {2025},
  eprint        = {2502.13189},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2502.13189},
  url           = {https://arxiv.org/abs/2502.13189}
}

@inproceedings{acharya2025starattention,
  title         = {{Star Attention}: Efficient {LLM} Inference over Long Sequences},
  author        = {Acharya, Shantanu and Jia, Fei and Ginsburg, Boris},
  booktitle     = {Proceedings of the 42nd International Conference on Machine Learning},
  series        = {Proceedings of Machine Learning Research},
  volume        = {267},
  pages         = {356--371},
  publisher     = {PMLR},
  year          = {2025},
  eprint        = {2411.17116},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2411.17116},
  url           = {https://proceedings.mlr.press/v267/acharya25a.html}
}

@inproceedings{xu2025xattention,
  title         = {{XAttention}: Block Sparse Attention with Antidiagonal Scoring},
  author        = {Xu, Ruyi and Xiao, Guangxuan and Huang, Haofeng and Guo, Junxian and Han, Song},
  booktitle     = {Proceedings of the 42nd International Conference on Machine Learning},
  series        = {Proceedings of Machine Learning Research},
  volume        = {267},
  pages         = {69819--69831},
  publisher     = {PMLR},
  year          = {2025},
  eprint        = {2503.16428},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2503.16428},
  url           = {https://arxiv.org/abs/2503.16428}
}

@article{deng2026unique,
  title         = {{UNIQUE}: Universal Top-K Sparse Attention for Training-Free Inference and Sparsity-Aware Training},
  author        = {Deng, Keqi and Ling, Shaoshi and Fan, Ruchao and Li, Jinyu},
  journal       = {arXiv preprint arXiv:2605.27740},
  year          = {2026},
  eprint        = {2605.27740},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2605.27740},
  url           = {https://arxiv.org/abs/2605.27740}
}

% ============================================================================
% 8. Adaptive and general-purpose sparse kernels
% ============================================================================

@inproceedings{zhang2025spargeattention,
  title         = {{SpargeAttention}: Accurate and Training-free Sparse Attention Accelerating Any Model Inference},
  author        = {Zhang, Jintao and Xiang, Chendong and Huang, Haofeng and Wei, Jia and Xi, Haocheng and Zhu, Jun and Chen, Jianfei},
  booktitle     = {Proceedings of the 42nd International Conference on Machine Learning},
  series        = {Proceedings of Machine Learning Research},
  volume        = {267},
  pages         = {76397--76413},
  publisher     = {PMLR},
  year          = {2025},
  eprint        = {2502.18137},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2502.18137},
  url           = {https://proceedings.mlr.press/v267/zhang25ch.html}
}

@inproceedings{goncalves2025adasplash,
  title         = {{AdaSplash}: Adaptive Sparse Flash Attention},
  author        = {Gon{\c{c}}alves, Nuno and Treviso, Marcos and Martins, Andr{\'e} F. T.},
  booktitle     = {Proceedings of the 42nd International Conference on Machine Learning},
  series        = {Proceedings of Machine Learning Research},
  volume        = {267},
  pages         = {19878--19896},
  publisher     = {PMLR},
  year          = {2025},
  eprint        = {2502.12082},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  doi           = {10.48550/arXiv.2502.12082},
  url           = {https://proceedings.mlr.press/v267/goncalves25a.html}
}

@article{zhang2026spargeattention2,
  title         = {{SpargeAttention2}: Trainable Sparse Attention via Hybrid Top-K+Top-P Masking and Distillation Fine-Tuning},
  author        = {Zhang, Jintao and Jiang, Kai and Xiang, Chendong and Feng, Weiqi and Hu, Yuezhou and Xi, Haocheng and Chen, Jianfei and Zhu, Jun},
  journal       = {arXiv preprint arXiv:2602.13515},
  year          = {2026},
  eprint        = {2602.13515},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  doi           = {10.48550/arXiv.2602.13515},
  url           = {https://arxiv.org/abs/2602.13515}
}

@article{goncalves2026adasplash2,
  title         = {{AdaSplash-2}: Faster Differentiable Sparse Attention},
  author        = {Gon{\c{c}}alves, Nuno and Pitorro, Hugo and Niculae, Vlad and Ponti, Edoardo and Li, Lei and Martins, Andr{\'e} F. T. and Treviso, Marcos},
  journal       = {arXiv preprint arXiv:2604.15180},
  year          = {2026},
  eprint        = {2604.15180},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2604.15180},
  url           = {https://arxiv.org/abs/2604.15180}
}

% ============================================================================
% 9. Vision and video-generation sparse attention
% ============================================================================

@inproceedings{wei2023sparsifiner,
  title     = {{Sparsifiner}: Learning Sparse Instance-Dependent Attention for Efficient Vision Transformers},
  author    = {Wei, Cong and Duke, Brendan and Jiang, Ruowei and Aarabi, Parham and Taylor, Graham W. and Shkurti, Florian},
  booktitle = {2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  pages     = {22680--22689},
  publisher = {IEEE},
  year      = {2023},
  doi       = {10.1109/CVPR52729.2023.02172},
  url       = {https://doi.org/10.1109/CVPR52729.2023.02172}
}

@article{liu2025fpsattention,
  title         = {{FPSAttention}: Training-Aware {FP8} and Sparsity Co-Design for Fast Video Diffusion},
  author        = {Liu, Akide and Zhang, Zeyu and Li, Zhexin and Bai, Xuehai and Han, Yizeng and Tang, Jiasheng and Xing, Yuanjie and Wu, Jichao and Yang, Mingyang and Chen, Weihua and He, Jiahao and He, Yuanyu and Wang, Fan and Haffari, Gholamreza and Zhuang, Bohan},
  journal       = {arXiv preprint arXiv:2506.04648},
  year          = {2025},
  eprint        = {2506.04648},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  doi           = {10.48550/arXiv.2506.04648},
  url           = {https://arxiv.org/abs/2506.04648}
}

@article{hu2026dfsattn,
  title         = {{DFSAttn}: Dynamic Fine-Grained Sparse Attention for Efficient Video Generation},
  author        = {Hu, Jie and Gao, Zixiang and He, Yutong and Yuan, Kun},
  journal       = {arXiv preprint arXiv:2605.23445},
  year          = {2026},
  eprint        = {2605.23445},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  doi           = {10.48550/arXiv.2605.23445},
  url           = {https://arxiv.org/abs/2605.23445}
}

@inproceedings{tan2026dsv,
  title         = {{DSV}: Exploiting Dynamic Sparsity to Accelerate Large-Scale Video {DiT} Training},
  author        = {Tan, Xin and Chen, Yuetao and Jiang, Yimin and Chen, Xing and Yan, Kun and Duan, Nan and Zhu, Yibo and Jiang, Daxin and Xu, Hong},
  booktitle     = {Proceedings of the 31st ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 1},
  pages         = {101--116},
  publisher     = {ACM},
  year          = {2026},
  eprint        = {2502.07590},
  archivePrefix = {arXiv},
  primaryClass  = {cs.DC},
  doi           = {10.1145/3760250.3762216},
  url           = {https://arxiv.org/abs/2502.07590}
}

@article{durvasula2025fgattn,
  title         = {{FG-Attn}: Leveraging Fine-Grained Sparse Attention in Video Diffusion Models},
  author        = {Durvasula, Sankeerth and Sreedhar, Kavya and Moustafa, Zain and Kothawade, Suraj and Pang, Tianlei and Gondimalla, Ashish and Subramanian, Suvinay and Shahidi, Narges and Vijaykumar, Nandita},
  journal       = {arXiv preprint arXiv:2509.16518},
  year          = {2025},
  eprint        = {2509.16518},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  doi           = {10.48550/arXiv.2509.16518},
  url           = {https://arxiv.org/abs/2509.16518}
}

@article{zhang2025vsa,
  title         = {{VSA}: Faster Video Diffusion with Trainable Sparse Attention},
  author        = {Zhang, Peiyuan and Chen, Yongqi and Huang, Haofeng and Lin, Will and Liu, Zhengzhong and Stoica, Ion and Xing, Eric and Zhang, Hao},
  journal       = {arXiv preprint arXiv:2505.13389},
  year          = {2025},
  eprint        = {2505.13389},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  doi           = {10.48550/arXiv.2505.13389},
  url           = {https://arxiv.org/abs/2505.13389}
}

@article{yang2025sparsevideogen2,
  title         = {{Sparse VideoGen2}: Accelerating Video Generation with Sparse Attention via Semantic-Aware Permutation},
  author        = {Yang, Shuo and Xi, Haocheng and Zhao, Yilong and Li, Muyang and Zhang, Jintao and Cai, Han and Lin, Yujun and Li, Xiuyu and Xu, Chenfeng and Chen, Jianfei and Han, Song and Keutzer, Kurt and Stoica, Ion},
  journal       = {arXiv preprint arXiv:2505.18875},
  year          = {2025},
  eprint        = {2505.18875},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  doi           = {10.48550/arXiv.2505.18875},
  url           = {https://arxiv.org/abs/2505.18875}
}

@article{chen2025dbsp,
  title         = {{db-SP}: Accelerating Sparse Attention for Visual Generative Models with Dual-Balanced Sequence Parallelism},
  author        = {Chen, Siqi and Hong, Ke and Zhao, Tianchen and Xie, Ruiqi and Zhu, Zhenhua and Zhang, Xudong and Wang, Yu},
  journal       = {arXiv preprint arXiv:2511.23113},
  year          = {2025},
  eprint        = {2511.23113},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  doi           = {10.48550/arXiv.2511.23113},
  url           = {https://arxiv.org/abs/2511.23113}
}

% ============================================================================
% 10. Exact dense-attention performance baselines
% ============================================================================

@inproceedings{dao2022flashattention,
  title         = {{FlashAttention}: Fast and Memory-Efficient Exact Attention with {IO}-Awareness},
  author        = {Dao, Tri and Fu, Daniel Y. and Ermon, Stefano and Rudra, Atri and R{\'e}, Christopher},
  booktitle     = {Advances in Neural Information Processing Systems},
  volume        = {35},
  pages         = {16344--16359},
  year          = {2022},
  eprint        = {2205.14135},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2205.14135},
  url           = {https://proceedings.neurips.cc/paper_files/paper/2022/hash/67d57c32e20fd0a7a302cb81d36e40d5-Abstract-Conference.html}
}

@inproceedings{dao2024flashattention2,
  title         = {{FlashAttention-2}: Faster Attention with Better Parallelism and Work Partitioning},
  author        = {Dao, Tri},
  booktitle     = {The Twelfth International Conference on Learning Representations},
  year          = {2024},
  eprint        = {2307.08691},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2307.08691},
  url           = {https://arxiv.org/abs/2307.08691}
}

@inproceedings{shah2024flashattention3,
  title         = {{FlashAttention-3}: Fast and Accurate Attention with Asynchrony and Low-Precision},
  author        = {Shah, Jay and Bikshandi, Ganesh and Zhang, Ying and Thakkar, Vijay and Ramani, Pradeep and Dao, Tri},
  booktitle     = {Advances in Neural Information Processing Systems},
  volume        = {37},
  year          = {2024},
  eprint        = {2407.08608},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2407.08608},
  url           = {https://proceedings.neurips.cc/paper_files/paper/2024/hash/7ede97c3e082c6df10a8d6103a2eebd2-Abstract-Conference.html}
}

@inproceedings{zadouri2026flashattention4,
  title         = {{FlashAttention-4}: Algorithm and Kernel Pipelining Co-Design for Asymmetric Hardware Scaling},
  author        = {Zadouri, Ted and Hoehnerbach, Markus and Shah, Jay and Liu, Timmy and Thakkar, Vijay and Dao, Tri},
  booktitle     = {Proceedings of Machine Learning and Systems},
  year          = {2026},
  eprint        = {2603.05451},
  archivePrefix = {arXiv},
  primaryClass  = {cs.LG},
  doi           = {10.48550/arXiv.2603.05451},
  url           = {https://arxiv.org/abs/2603.05451}
}