The stoat project is a comprehensive initiative designed to enhance the development of essential, foundational, and practical statistical toolkits. Its overarching aim is to widen the scope of operability and practicality intrinsic to statistical concepts and methodologies. The principles of the Stoat project can be underscored as follows:
- Open. stoat supports open-source, and all toolkits will be published in Github.
- Decentralization. stoat does not claim credits for every project or toolkit. Rather, all credits associated with a particular project are duly accorded to the respective developers and contributors.
- Cooperation. stoat involves active cooperation between developers/contributors and mentors. Most projects are initiated by mentors who propose preliminary ideas and aims. The project then is progressively proceeded through collaboration between contributors and mentors.
- Paid-position. stoat offers paid short-term academic positions offered by CUHK-STAT, such as student helper or research assistant (RA). In short, each contributor selected for a stoat project will get paid academic position to work on an Python/R package for 12 - 24 weeks.
| Date | Proposal | Github | Status | languages | Contributors |
|---|---|---|---|---|---|
| 10/25 | Poisson Binomial Distribution for PyTorch | Hiring | Python | ||
| 08/25 | Scikit-learn Compatible plqERM Estimator | Ongoing | Python | Youtong LI |
We Support Your Proposal. If you have clear objectives and strategies for a statistical toolkit, you are invited to upload your project proposal (see Proposal template) in the discussion section under project proposal category. PIs will assess its feasibility and provide support accordingly.
| Date | Proposal | Outcome | languages | Contributors |
|---|---|---|---|---|
| 12/23 | PLQ Composite Decomposition | Github | Python | Tingxian Gao |
| 03/24 | Portfolio Optimization via ReHLine (pre) | Github | Python | Alibek Orazalin |
| 01/25 | Fast Path Solution for plqERM | Github | Python | Youtong LI |
| 07/25 | Fast Path Solution for CQR | Github | Python | Youtong LI |
| 01/25 | Matrix Factorization Optimization with various Loss functions | Github | Python | Xiaochen Su |
Student who wants to participate in SToAT project should:
- please review the README of the SToAT project to ensure that it aligns well with your expertise and interests;
- please check and follow the SToAT-wiki: Participation workflow which provides an in-depth explanation of the workflow for the SToAT project; also consider referring to SToAT-wiki: application template;
- correspond with mentors via email to discuss the relevant project, if necessary.
stoat is initiated by several junior statistics PIs to contribute to the development of basic, fundamental and practical statistical toolkits. stoat is initially inspired by Google Summer of Code initiative, and provides more flexible timeline, research-oriented projects, and offers paid university-based academic job positions.
Support SToAT. Stoat enthusiastically welcomes more co-PIs to join, providing opportunities for students interested in statistical software development, as well as contributing to the development and promotion of open-source statistical toolkits. The role of the PI is pivotal in overseeing the project, providing vision, direction, and coordinating the contributors' efforts. Please check SToAT-wiki: Become a co-PI of SToAT for more information.