GitHub project reproduces 58 Schmidhuber papers with AI assistant
Yaroslavvb's GitHub project uses an AI coding assistant to reproduce 58 implementations of Jürgen Schmidhuber's papers from 1990 to 2025. The repository features exact recreations of original algorithms including LSTM, World Models with VAE and RNN components, and Neural Data Router from ICLR 2022. Animated GIFs visualize training dynamics, forming a visual timeline of 35 years of research progression.
#18hardmaru@HARDMARUThe agent also reimplemented the “Blues Improvisation” experiment by @douglas_eck and @SchmidhuberAI in 2002 which show that LSTMs can learn temporal structure in music.
Finding temporal structure in music: Blues improvisation with LSTM recurrent networks https://sferics.idsia.ch/pub/juergen/2002_ieee.pdf
hardmaru@hardmaruReproducing all of Schmidhuber’s papers (1990-2025) using an AI coding assistant. Cool project by @yaroslavvb! It even reproduced the “World Models” paper by me and @SchmidhuberAI with a toy env, with a full VAE + RNN world model implementation. Project: https://github.com/cybertronai/schmidhuber-problems/blob/main/VISUAL_TOUR.md
#237Max Jaderberg@MAXJADERBERGThis project is such a cool timeline of progression!
hardmaru@hardmaruReproducing all of Schmidhuber’s papers (1990-2025) using an AI coding assistant. Cool project by @yaroslavvb! It even reproduced the “World Models” paper by me and @SchmidhuberAI with a toy env, with a full VAE + RNN world model implementation. Project: https://github.com/cybertronai/schmidhuber-problems/blob/main/VISUAL_TOUR.md
@hardmaru @SchmidhuberAI I'm sure the agent wrote better code than my 50k+ lines of C++ in 2000; I hope it likes blues music.
hardmaru@hardmaruThe agent also reimplemented the “Blues Improvisation” experiment by @douglas_eck and @SchmidhuberAI in 2002 which show that LSTMs can learn temporal structure in music. Finding temporal structure in music: Blues improvisation with LSTM recurrent networks https://sferics.idsia.ch/pub/juergen/2002_ieee.pdf
everything is reproducible now
hardmaru@hardmaruReproducing all of Schmidhuber’s papers (1990-2025) using an AI coding assistant. Cool project by @yaroslavvb! It even reproduced the “World Models” paper by me and @SchmidhuberAI with a toy env, with a full VAE + RNN world model implementation. Project: https://github.com/cybertronai/schmidhuber-problems/blob/main/VISUAL_TOUR.md