Divide and Generate: Neural Generation of Complex Sentences
This work addresses a domain-specific task in natural language generation, offering an incremental improvement for text generation systems.
The paper tackles the problem of generating complex sentences from simple ones to enhance database responses, achieving superior performance with a pipeline model over an end-to-end model in automatic evaluations.
We propose a task to generate a complex sentence from a simple sentence in order to amplify various kinds of responses in the database. We first divide a complex sentence into a main clause and a subordinate clause to learn a generator model of modifiers, and then use the model to generate a modifier clause to create a complex sentence from a simple sentence. We present an automatic evaluation metric to estimate the quality of the models and show that a pipeline model outperforms an end-to-end model.