We propose to tackle data-to-text generation tasks by directly splicing together retrieved segments of text from neighbor'' source-target pairs. Unlike recent work that conditions on retrieved neighbors but generates text token-by-token, left-to-righ
t, we learn a policy that directly manipulates segments of neighbor text, by inserting or replacing them in partially constructed generations. Standard techniques for training such a policy require an oracle derivation for each generation, and we prove that finding the shortest such derivation can be reduced to parsing under a particular weighted context-free grammar. We find that policies learned in this way perform on par with strong baselines in terms of automatic and human evaluation, but allow for more interpretable and controllable generation.
