Awide range ofreinforcement learning (RL) problems --including robustness, transfer learning, unsupervised RL, and emergent complexity -- require specifying a distribution of tasks or environments in which a policy will be trained.

Weformulate the decoding process asanoptimization problem which allows for multiple attributesweaimtocontrol tobeeasilyincorporated asdifferentiable constraints to the optimization. By relaxing this discrete optimization to a continuous one, we make use of Lagrangian multipliers and gradient-descent based techniques to generate the desired text.

Stochastic approximation is a foundation for many algorithms found in machine learning and optimization.