British artificial intelligence specialist DeepMind Technologies, acquired by Google parent company Alphabet in 2014, has announced the release of a library of building blocks for reinforcement learning in machine intelligence: TRFL.

The TRFL library, pronounced “truffle”, is described by the company as “a collection of key algorithmic components that we have used internally for a large number of our most successful agents such as DQN, DDPG and the Importance Weighted Actor Learner Architecture.” The library includes functions for the implementation of classic reinforcement learning algorithms as well as what the company claims are “more cutting-edge techniques,” both implemented in Google’s open-source TensorFlow language.

“They are not complete algorithms,” the DeepMind team warns, “but implementations of RL-specific mathematical operations needed when building fully-functional RL agents. This is not a one-time release. Since this library is used extensively within DeepMind, we will continue to maintain it as well as add new functionalities over time. We are also eager to receive contributions to the library by the wider RL community.”

The TRFL library is available now from DeepMind’s GitHub repository, under an Apache v2.0 licence.