exAClior / NumericTrainingCamp

Training camp at HKUST-GZ, about numeric methods in Physics

Scientific Computing Training Camp (Basic)

May 27 to June 9, 2024 (14 days) at Hong Kong University of Science and Technology (Guangzhou). Not open for registration.

During the 14 days, you will implement an algorithm, simulate a many-body system, and write a report.

Talk is cheap. Show me the code. - Linus Torvalds

Topics

1. Tensor Networks
2. AI for Science
3. Quantum Monte Carlo
4. Computational Complexity and Quantum Computing

Inputs

1. Work on a project under the chosen topic, in the meeting room, from 9:30AM-9:30PM.
2. Bidaily presentation and report at 4:00 PM, each group has 15min to report the project progress in the meeting. Write a report and submit it to this GitHub repo before 9:30PM. The report should be written in Markdown format.
3. A final report in the Markdown format, which will be deployed to the training camp website.

Note: each participant (including instructors) has 48 hour leave. Quitting is only allowed at the first 3 days.

Expected output

1. A final report in the Markdown format.
2. A code repository.
3. Some research projects (not guaranteed).

Events

Mini-lectures

1. Jin-Guo Liu, Tensor network methods, May 27, 9:30AM
2. Ze-Cheng Gan, Differential equation models, algorithms and applications in Molecular Dynamics, May 29, 9:30AM
3. Qi Zhang, May 30, 9:30AM, Flow model for computational physics¹ (GitHub repo: https://github.com/zhangqi94/Flow)

Talks

1. Yi-Jia Wang, June 3, 11AM, online, Tensor network message passing²
2. Ji-Yao Chen, June 5, 11AM, W4-202, Tensor networks, from ground states to excitations and back
3. Xiu-Hao Deng, Juan 6, 2PM, Quantum random walk (planned)
4. Zhao-Long Gu, June 8, 11AM, Correlation Effects and Their Interplay with Magnetic Fields and Band Topology in Twisted Transition Metal Dichalcogenides

Guide

Conduct a survey

You can use Google Scholar to search for the researchers you are interested in, and follow them to get updated with the latest research. You can use Zotero to manage your references. For more details, please refer to this guide.

Technical writing

The report in the Markdown format should be submitted to this repository, under the folder of the corresponding topic. Please check preview events (e.g. this one) for references.

• Guide: How to use AI to facilitate technical writing

How to program?

• You must use git to manage your code, and use pull requests to submit your code. Please refer to this guide for more details.
• Code (Agile software development), even if you do not understand the algorithm at all.
• Remove uncertainty, you can get help from me, or the community, e.g. https://julialang.org/community/
  • I want to ..., but I don't know how to start.
  • Is there any good example/existing code about...
  • Is it possible to help me review my code? Ref: Blog: The Best Way to Do a Code Review on GitHub

Logistics

• Spaces that free to use: W1 4F-6F, maker space.
• We use Zulip@HKUST-GZ stream: #NumericTrainingCamp for (formal) online communication.
• We will have a WeChat group for informal communication (planned).

Technical support

• Yusheng ZHAO
• Yidai ZHANG

References

Footnotes

1. Machine learning for physicists, https://github.com/wangleiphy/ml4p ↩

2. Wang, Y., Zhang, Y.E., Pan, F., Zhang, P., 2024. Tensor Network Message Passing. Phys. Rev. Lett. 132, 117401. https://doi.org/10.1103/PhysRevLett.132.117401 ↩