• Computer Science
  • Published in NAACL-HLT 2019

Step-by-Step: Separating Planning from Realization in Neural Data-to-Text Generation

@inproceedings{Moryossef2019StepbyStepSP,
  title={Step-by-Step: Separating Planning from Realization in Neural Data-to-Text Generation},
  author={Amit Moryossef and Yoav Goldberg and Ido Dagan},
  booktitle={NAACL-HLT},
  year={2019}
}
Data-to-text generation can be conceptually divided into two parts: ordering and structuring the information (planning), and generating fluent language describing the information (realization). Modern neural generation systems conflate these two steps into a single end-to-end differentiable system. We propose to split the generation process into a symbolic text-planning stage that is faithful to the input, followed by a neural generation stage that focuses only on realization. For training a… CONTINUE READING

Citations

Publications citing this paper.
SHOWING 1-10 OF 19 CITATIONS

Paraphrase Generation with Latent Bag of Words

VIEW 5 EXCERPTS
CITES BACKGROUND & METHODS
HIGHLY INFLUENCED

Paraphrase Generation with Latent Bag of Words

VIEW 5 EXCERPTS
CITES METHODS & BACKGROUND
HIGHLY INFLUENCED

Paraphrase Generation with Latent Bag of Words

VIEW 5 EXCERPTS
CITES BACKGROUND & METHODS
HIGHLY INFLUENCED

References

Publications referenced by this paper.
SHOWING 1-10 OF 39 REFERENCES

Pointing the Unknown Words

VIEW 5 EXCERPTS
HIGHLY INFLUENTIAL

Findings of the E2E NLG Challenge

VIEW 4 EXCERPTS
HIGHLY INFLUENTIAL

Challenges in Data-to-Document Generation

VIEW 3 EXCERPTS
HIGHLY INFLUENTIAL