MATS Winter 2023-24 Retrospective
webAuthor
Credibility Rating
Good quality. Reputable source with community review or editorial standards, but less rigorous than peer-reviewed venues.
Rating inherited from publication venue: EA Forum
MATS is a prominent AI safety research mentorship program; this retrospective is useful for understanding field-building strategies, program evaluation methods, and the talent pipeline entering AI safety research.
Forum Post Details
Metadata
Summary
Comprehensive retrospective on the fifth iteration of the ML Alignment & Theory Scholars (MATS) program, covering 63 scholars mentored by 20 AI safety researchers. The report evaluates program outcomes, scholar satisfaction, and operational changes, while identifying improvements for future cohorts including advisory boards for mentor selection and expanded AI governance mentorship.
Key Points
- •63 scholars received mentorship from 20 research mentors; scholars rated overall program 9.2/10 likelihood to recommend (NPS +74)
- •Key operational changes included reducing stipends from $40/h to $30/h, transitioning to a Research Management model, and expanding use of Lighthaven campus
- •Mentorship was the most valuable element for 38% of scholars, rated 8.1/10; median scholar made 5 professional connections and found 5 potential collaborators
- •Research Management added ~10% productivity boost and 10 productive hours per median scholar, valued at ~$1000
- •Future improvements proposed: advisory boards for mentor selection, fewer mechanistic interpretability mentors, more AI governance mentors, and pre-screening with CodeSignal
Cited by 1 page
| Page | Type | Quality |
|---|---|---|
| MATS ML Alignment Theory Scholars program | Organization | 60.0 |
Cached Content Preview
# MATS Winter 2023-24 Retrospective
By utilistrutil
Published: 2024-05-11
Co-Authors: [@Rocket](https://www.lesswrong.com/users/rocket?mention=user), [@LauraVaughan](https://www.lesswrong.com/users/lauravaughan?mention=user), [@McKennaFitzgerald](https://www.lesswrong.com/users/mckennafitzgerald?mention=user), [@Christian Smith](https://www.lesswrong.com/users/christian-smith?mention=user), [@Juan Gil](https://www.lesswrong.com/users/juan-gil?mention=user), [@Henry Sleight](https://www.lesswrong.com/users/henry-sleight?mention=user), [@Matthew Wearden](https://www.lesswrong.com/users/matthew-wearden?mention=user), [@Ryan Kidd](https://www.lesswrong.com/users/ryankidd44?mention=user)
The [ML Alignment & Theory Scholars](https://www.matsprogram.org) program (MATS) is an education and research mentorship program for researchers entering the field of AI safety. This winter, we held the fifth iteration of the MATS program, in which 63 scholars received mentorship from 20 research mentors. In this post, we motivate and explain the elements of the program, evaluate our impact, and identify areas for improving future programs.
Summary
=======
Key details about the Winter Program:
* The four main changes we made after our Summer program were:
* Reducing our scholar stipend from $40/h to $30/h based on alumni feedback;
* Transitioning [Scholar Support](https://www.lesswrong.com/posts/zwf68YaySvXhWYCdh/mats-summer-2023-retrospective#Scholar_Support) to [Research Management](https://www.lesswrong.com/posts/Z87fSrxQb4yLXKcTk/mats-winter-2023-24-retrospective#Research_Management);
* Using the full Lighthaven campus for office space as well as housing;
* Replacing [Alignment 201](https://course.aisafetyfundamentals.com/alignment-201) with [AI Strategy Discussions](https://www.lesswrong.com/posts/JsjJuikJsidkyfhyr/mats-ai-safety-strategy-curriculum).
* Educational attainment of MATS scholars:
* 48% of scholars were pursuing a bachelor’s degree, master’s degree, or PhD;
* 17% of scholars had a master’s degree as their highest level of education;
* 10% of scholars had a PhD.
* If not for MATS, scholars might have spent their counterfactual winters on the following pursuits (multiple responses allowed):
* Conducting independent alignment research without mentor (24%);
* Working at a non-alignment tech company (21%);
* Conducting independent alignment research with a mentor (13%);
* Taking classes (13%).
Key takeaways from scholar impact evaluation:
* Scholars are highly likely to recommend MATS to a friend or colleague (average likelihood is 9.2/10 and [NPS](https://en.wikipedia.org/wiki/Net_promoter_score) is +74).
* Scholars rated the mentorship they received highly (average rating is 8.1/10).
* For 38% of scholars, mentorship was the most valuable element of MATS.
* Scholars are likely to recommend Research Management to future scholars (average likelihood is 7.9/10 and [NP
... (truncated, 98 KB total)705ccf8dae23a800 | Stable ID: sid_vT6f1ApcSz