25launches

Energy-Based Reasoners

A 1.8M-parameter model that cracks the hardest Sudoku

0 followers

Visit energy-based-reasoners.github.io (opens in a new tab)

Artificial IntelligenceOpen Source

Energy-Based Reasoners (EBRs) are a research paper with code from Duke, UIUC and Harvard. Instead of looping a network a fixed number of times, an EBR learns an energy score for its current reasoning state, refines it by gradient descent, stops when it settles and keeps the lowest-energy answer. With 1.8M parameters it solves 99.4% of Sudoku-Extreme in one try and 99.9% with a few samples, using 345x less inference compute than the next-best model. Code is MIT; training needs a CUDA GPU.