Hi!
This is Yang Qizhen from Shanghai, China.
I speak Shanghainese, Mandarin, and English.
I am interested in linguistics, computer science, natural language processing (NLP), cryptocurrency, mathematics, anthropology, typography, ultimate frisbee, and visual arts.
Not only that, I am also a big fan of self-hosted and corp-free services. I believe we should all expel intrusive corporation products that do not act in favour of us. Everyone should live a sustainable and private digital life and enjoy the right to be forgotten.
PUBLIC KEY
(3992 5D88 A147 656C 8EAA 5261 B782 D7F5 7541 8B17)
-----BEGIN PGP PUBLIC KEY BLOCK----- mQENBGh6Hu8BCACxkASo7RMmERWWWBd3G2KoejTLnRVX3vNXerLbbzbWwsW1iPz8 3aNHyGOoRvpY5ufC1CLsbl+Xr6FHcrG2avuxG/BihGuOc4Q3t2IRb1P1Q6DBG2Mo +83x1sOq3vOC5NWjSt+3S/I3rj9eCoZzBAU6P2fgAw7Do+VumbgRToc4DFoDecWZ bop+FjoP/5UvmK61PQ95zTwQYZDG9iT+wNErmBr0AsmhXh53Zc2HaWxS33Q4kf34 DAxrSydx0AatbnXh62rGhlIX9gFir3XmmiHkG6D2l3QY0yMpsxKyw0VC8iDjvpyt v9ZP5n06VNR+CGC9S7dAdFzT/gONZvYk1FhnABEBAAG0L1lhbmcgQ2h5aSBKZW4g KHExemhlbikgPHFpemhlbnlhbmdzaEBnbWFpbC5jb20+iQFRBBMBCgA7FiEEOZJd iKFHZWyOqlJht4LX9XVBixcFAmh6Hu8CGwMFCwkIBwICIgIGFQoJCAsCBBYCAwEC HgcCF4AACgkQt4LX9XVBixcKQAf/XBE3PUnCuv1EVzVpuW7kJJTlpDNHQahcaKdb vSvRTQMpNU5jqAG0A3deWmv55um1LHGB4nArA2/y1dakmYh8/68XSmSiSs9gV5kb mIZd6r81XS/7FUmfYudxPu/rARLcxEm5nfdf3o1VGvCiCuawt0YMD3qttm1F+6Fb /r6dSMyrdRfzJanprnDXdcDO+Jjwmg3JN377NumnI9cBaoErA0MbYeyqdytSbf7z +P1i4rnnqEneQJUCwoe+jd3m1Eq1U7l+puYYqQWCQqXPmt/nH8HPyM/mzenRUr8f YkawBTkLNRI2G2P/8PyWk5KWg6wQqeCmkUosmaJZKI1QXTv3YLkBDQRoeh7vAQgA 5MWiuXb6nBFtneDIyU1ScxfZbQgfVWtN9ot4Z6dmaUEAnlFQkwG5zpjuSgxhtPZO Eu3+9s4R582rCCf1FyDyTCrihrYrV2hdJMpKR9kz5KatFqLvUk+dRERJkjfpsKn8 nik66rk8/MUEZ+BGHIth+mDxviVZrNa1pCS+urGZQlCzO9UegV6rjE+rIsXHnoRx wNKrv/BOGZ8W2nByNmztwTan9g2VAKxwqQh3DVUheDL5wNMsUCCd1BZG3rGqyfHT l8gd/rdoC6bSMzs6TRP4+LTxRqYjv1UJu7T9cpJrQwwYHfzAUL15QgZsEJjy2MqX bvNpJYJOd/e9gNoRRE5KrQARAQABiQE2BBgBCgAgFiEEOZJdiKFHZWyOqlJht4LX 9XVBixcFAmh6Hu8CGwwACgkQt4LX9XVBixcpkAf/W9cmpw8p0GAJw6Yv1b0fwi8g boch+7MwxhhWxXhWOkYrR8lWgTa7VdID661KvhmhVRkZbBbdVwtsGjXhcw0mZxZT BdacRWkKfwZiHCkVc1DIaDRVxHuF6LKDGTU5WY2jQvdZPqH6PUnA2akD57ylCJGF 8kFTr3X84DXV+U9v5ZZK8tyxaqe1HDLvmVcvMoEK6EGR0bMTILUgfNzVErd9uGcR gakVQUmpzJa/eL8MROypw1rGxUe1nn23nyL8BYoEG2sX3LdN1p1wBg5g6SmNKVqE 5DI6Zu/EsoK/IbfX167H0Gjlaq5y5TYOaO+zH74EJ6BjDCldPVlM/qpMqhMtFw== =dBzH -----END PGP PUBLIC KEY BLOCK-----
Latest Blogs
Linux on Apple Silicon
·
linux-stuff
Just got a M2 Macbook Air and installed Fedora Asahi Remix on it.
Thoughts on altcoins
·
cryptocurrency
Will we see an altseason in this cycle?
More Blogs »
Research and Publications
– Present
ShUD: the First Shanghainese Universal Dependency Treebank
Read Paper »
Treebank included in the Universal Dependencies project; Accepted paper in UD Workshop 2025 (part of SyntaxFest 2025, an ACL-affiliated conference, in Ljubljana, 26–29 August 2025)
Treebank included in the Universal Dependencies project; Accepted paper in UD Workshop 2025 (part of SyntaxFest 2025, an ACL-affiliated conference, in Ljubljana, 26–29 August 2025)
Abstract
This paper introduces ShUD, the first Universal Dependencies (UD) treebank for Shanghainese, a Wu Chinese variant spoken by approximately 14 million people but severely under-resourced in NLP. The treebank is built through a scalable annotation pipeline that exploits grammatical parallels between Shanghainese and Mandarin. Our pipeline also provides a practical strategy for bootstrapping resources for other Chinese dialects. We documented syntactic phenomena unique to Shanghainese within the UD framework and fine-tuned a dependency parser using our annotated treebank, contributing a foundation to both NLP tool development and cross-linguistic syntactic research.
July 2025
Enhancing Fluid Dynamics Simulations: Integrating Neural Operators into Visual Models
Published in 2025 IEEE 7th International Conference on Artificial Intelligence, Computer Science, and Information Processing (AICSIP), in Hangzhou, 25–26 July 2025
Abstract
Fluid dynamics simulations, a cornerstone for lots of scientific fields, often require solving partial differential equations (PDEs), e.g. Navier-Strokes equations. Traditional methods such as computational fluid dynamics (CFD) are typically computationally intensive and have limitations on real-time simulations and extreme boundary conditions. This research discovers the potential of U-Net-like deep learning architecture, as a visual model, to simulate fluid motions and integrates neural operators, particularly Fourier Neural Operator (FNO) and Adaptive Fourier Neural Operator (AFNO), into its architecture to enhance and specialize its capability in capturing features in fluid motions and provides an alternative approach. Experiments are done on shallow water equations (single property) and Navier-Strokes equations (multiple properties). The modified architecture showcases a significant improvement compared with the baseline model, especially in single-property simulations and shows promise in multi-property scenarios, with improvements up to 80% in the error.