Discovering Textual Structures: Generative Grammar Induction using Template Trees
This provides a stepping stone for human-machine co-creation of generative models in applications like summaries and chatbots, though it is incremental as it builds on existing grammar and template methods.
The paper tackles the problem of learning interpretable grammars for natural language generation by introducing Gitta, a grammar induction algorithm that uses template trees to approximate human-created grammars with only a few examples, achieving reasonable approximation.
Natural language generation provides designers with methods for automatically generating text, e.g. for creating summaries, chatbots and game content. In practise, text generators are often either learned and hard to interpret, or created by hand using techniques such as grammars and templates. In this paper, we introduce a novel grammar induction algorithm for learning interpretable grammars for generative purposes, called Gitta. We also introduce the novel notion of template trees to discover latent templates in corpora to derive these generative grammars. By using existing human-created grammars, we found that the algorithm can reasonably approximate these grammars using only a few examples. These results indicate that Gitta could be used to automatically learn interpretable and easily modifiable grammars, and thus provide a stepping stone for human-machine co-creation of generative models.