sigmoid.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A social space for people researching, working with, or just interested in AI!

Server stats:

654
active users

#probability

2 posts2 participants0 posts today

Over the past couple of years, I've really fallen in love with #tikz and all of its quirks.

TikZ is a plotting/graphics package for LaTeX that is especially useful for creating mathematical diagrams.

The support for mathematical notation is unbeatable and the flexibility of the language is extremely high. Also, graphics rendered to pdf/svg in this way are extremely lightweight and reproducible.

I do find it very challenging syntax to remember though, so I put together this GitHub repository to keep track of tikz code I've written.

github.com/ctesta01/tikz-examp

Each graphic shown in the README is linked to its underlying .tex code.

Also the README has several links to documentation / tutorials that I've found helpful along with some tips I've learned from experience.

Dear LazyWeb: is there a C/C++, #RustLang or #Zig equivalent of #SciPy’s `stats` module for statistical analysis? Namely:
• a collection of common PDFs (probability density functions);
• MLE (maximum likelihood estimation) for these common distributions;
• KDE (kernel density estimation).

SciPy’s API is a pleasure to work with. Anything that comes close but usable from C/C++/Rust/Zig would make my life so much easier. Boosts appreciated for visibility.

[2502.05244] Probabilistic Artificial Intelligence
arxiv.org/abs/2502.05244
news.ycombinator.com/item?id=4

Manuscript 418pp ...

arXiv.orgProbabilistic Artificial IntelligenceArtificial intelligence commonly refers to the science and engineering of artificial systems that can carry out tasks generally associated with requiring aspects of human intelligence, such as playing games, translating languages, and driving cars. In recent years, there have been exciting advances in learning-based, data-driven approaches towards AI, and machine learning and deep learning have enabled computer systems to perceive the world in unprecedented ways. Reinforcement learning has enabled breakthroughs in complex games such as Go and challenging robotics tasks such as quadrupedal locomotion. A key aspect of intelligence is to not only make predictions, but reason about the uncertainty in these predictions, and to consider this uncertainty when making decisions. This is what this manuscript on "Probabilistic Artificial Intelligence" is about. The first part covers probabilistic approaches to machine learning. We discuss the differentiation between "epistemic" uncertainty due to lack of data and "aleatoric" uncertainty, which is irreducible and stems, e.g., from noisy observations and outcomes. We discuss concrete approaches towards probabilistic inference and modern approaches to efficient approximate inference. The second part of the manuscript is about taking uncertainty into account in sequential decision tasks. We consider active learning and Bayesian optimization -- approaches that collect data by proposing experiments that are informative for reducing the epistemic uncertainty. We then consider reinforcement learning and modern deep RL approaches that use neural network function approximation. We close by discussing modern approaches in model-based RL, which harness epistemic and aleatoric uncertainty to guide exploration, while also reasoning about safety.