Deep Reinforcement Learning in Large Discrete Action Spaces
This addresses a bottleneck for applying reinforcement learning to real-world tasks with many discrete actions, representing an incremental improvement over existing methods.
The paper tackles the problem of applying reinforcement learning to environments with large discrete action spaces, such as recommender systems and language models, by proposing an approach that embeds actions in a continuous space and uses approximate nearest-neighbor methods for logarithmic-time lookup, enabling training on tasks with up to one million actions.
Being able to reason in an environment with a large number of discrete actions is essential to bringing reinforcement learning to a larger class of problems. Recommender systems, industrial plants and language models are only some of the many real-world tasks involving large numbers of discrete actions for which current methods are difficult or even often impossible to apply. An ability to generalize over the set of actions as well as sub-linear complexity relative to the size of the set are both necessary to handle such tasks. Current approaches are not able to provide both of these, which motivates the work in this paper. Our proposed approach leverages prior information about the actions to embed them in a continuous space upon which it can generalize. Additionally, approximate nearest-neighbor methods allow for logarithmic-time lookup complexity relative to the number of actions, which is necessary for time-wise tractable training. This combined approach allows reinforcement learning methods to be applied to large-scale learning problems previously intractable with current methods. We demonstrate our algorithm's abilities on a series of tasks having up to one million actions.