3 ms·
https://ulvgard.se https://ulvgard.se Reverse engineering algorithms and results from papers. I love the process of walking backwards from the results and figu
by tulvgard 5y ago
https://ulvgard.se https://ulvgard.se
Reverse engineering algorithms and results from papers. I love the process of walking backwards from the results and figure out how they were achieved, uncovering what the authors glanced over or failed to mention at all.
Two mentions:
- Facebook Prophet: https://ulvgard.se/articles/trend_and_seasonality_decomposition_with_prophet/ https://ulvgard.se/articles/trend_and_seasonality_decomposit...
- One-shot neural network training using hypercubes: https://ulvgard.se/articles/one_shot_training_neural_networks_using_geometric_bounds/ https://ulvgard.se/articles/one_shot_training_neural_network...
- ciprian_craciun 5y agoVery nice look-and-feel! It resembles perfectly reading a scientific paper (but without double columns and zig-zag scrolling through PDF)! The text is perhaps a little bit too small by default, but nothing that can't be solved with some zooming. The only downside is that formula rendering requires JavaScript. :( (I would love to see a proper, and simple solution, to rendering mathematical formulas without JavaScript...)
- tulvgard 5y agoThank you for the feedback! - Text is now increased slightly, color darkened for improved readability. Much appreciate you taking time to comment.