Leela (software)

Leela is the name given to several distinct open-source software projects, primarily focusing on artificial intelligence and game-playing. The most well-known usage refers to:

Leela Chess Zero (LCZero or Lc0): A free, open-source chess engine developed by Gary Linscott and contributors. It is notable for being trained primarily through reinforcement learning, specifically a neural network trained using self-play. This approach, similar to that employed by AlphaZero, allows Leela Chess Zero to learn and improve its chess-playing abilities without relying on human-engineered heuristics or opening books to the same extent as traditional chess engines. LCZero uses the Monte Carlo Tree Search (MCTS) algorithm for move selection, guiding its search based on the neural network's evaluation of board positions. Its distributed training model allows individuals to contribute computing power to the neural network's learning process. LCZero has demonstrated strong chess performance, achieving grandmaster-level play and competing successfully against other top chess engines.

Leela Zero: The broader project that encompasses Leela Chess Zero. It was originally conceived as an attempt to replicate AlphaGo's approach to playing the game of Go. Leela Zero for Go was also developed and showed promising results.

In summary, "Leela" often refers to the Leela Chess Zero chess engine, which is a notable example of artificial intelligence applied to game-playing, using reinforcement learning and distributed computing to achieve high levels of performance. However, it's important to remember that the "Leela" name originates from a larger AI project aimed at replicating AlphaGo's success in Go.