@article{graphkernelattentiontransformers, title = {From block-Toeplitz matrices to differential equations on graphs: towards a general theory for scalable masked Transformers}, author = {Krzysztof Choromanski and Han Lin and Haoxian Chen and Tianyi Zhang and Arijit Sehanobish and Valerii Likhosherstov and Jack Parker-Holder and Tamas Sarlos and Adrian Weller and Thomas Weingarten}, year = {2021}, eprint = {2107.07999}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2107.07999v8}, }