As we’ve been working on Squiggle, we’ve encountered many technical challenges for writing probabilistic functionality with Javascript. Some of these challenges are solved in Python and need to be ported over and some apply to all languages.
We think the following tasks could be good fits for others to tackle. These are fairly isolated and could be done in contained NPM packages or similar. The solutions would be useful for Squiggle and we suspect might be handy for others in the Javascript ecosystem as well. Advice and opinions are also appreciated.
This post was quickly written, as it's for a narrow audience and might get outdated. We're happy to provide more rigor and context if requested. If you are interested in taking any of them on and could use some guidance, let us know!
For those not themselves interested in contributing, this might be useful for giving people a better idea of [...]
Outline:
(01:00) 1. Density Estimation
(03:31) 2. Quantiles to Distributions, Maybe with Metalog
(04:22) 3. “Point Set” Formats
(08:39) 4. Scoring Heuristics, for General Point Set Distributions
(10:18) 5. D3 Scale Axes
(10:55) 6. Forced Distribution Correlations
(11:22) 7. Language Model Training
(11:40) 8. Distribution-Specific Tasks
(12:21) 9. Bonus: Relative Value Functions
First published:
August 26th, 2023
Source:
https://forum.effectivealtruism.org/posts/7aieQzPcBKxMGn9Xo/open-technical-challenges-around-probabilistic-programs-and
Narrated by TYPE III AUDIO.