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

By Amit Moryossef, Yoav Goldberg, Ido Dagan. 2019

198 citations recorded by the source corpus.

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 plan-to-text generator, we present a method for matching reference texts to their corresponding text plans. For inference time, we describe a method for selecting high-quality text plans for new inputs. We implement and evaluate our approach on the WebNLG benchmark. Our results demonstrate that decoupling text planning from neural realization indeed improves the system’s reliability and adequacy while maintaining fluent output. We observe improvements both in BLEU scores and in manual evaluations. Another benefit of our approach is the ability to output diverse realizations of the same input, paving the way to explicit control over the generated text structure.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Natural Language Processing

Citations in the corpus

PEER: A Collaborative Language Model (2022, background): “, 2021), data-to-text generation (Moryossef et al., 2019) and story writing (Yao et al.”
Survey of Hallucination in Natural Language Generation (2022, methodology): “Planning or skeleton is a common method in data-to-text tasks to improve the faithfulness to the input [124].”
DART: Open-Domain Structured Data Record to Text Generation (2020, methodology): “Furthermore, we also tried BestPlan (Moryossef et al., 2019b) and NeuralPlan (Moryossef et al., 2019a) models, both of which use a step-by-step approach with a text-planning stage followed by a plan-realization stage on the WebNLG dataset.”
Text-to-Text Pre-Training for Data-to-Text Tasks (2020, background): “Ferreira et al. (2019) and Moryossef et al. (2019) ﬁnd that neural pipelined approaches perform better than end-to-end models.”
Few-Shot NLG with Pre-Trained Language Model (2019, methodology): “…been used, including the end-to-end methods that jointly model context selection and surface realization (Liu et al. 2018; Wiseman, Shieber, and Rush 2018; Puduppully, Dong, and Lapata 2018), and the methods using neural network to tackle realization component (Moryossef, Goldberg, and Dagan 2019).”
Constrained Decoding for Neural NLG from Compositional Representations in Task-Oriented Dialogue (2019, background): “A more recent work by Moryossef et al. (2019) also focuses on exercising more control over input structures through sentence plans; however, their work doesn’t touch on discourse relations or constrained decoding.”
Investigating Pretrained Language Models for Graph-to-Text Generation (2020, methodology): “proaches for KG-to-text generation linearize the KG triples ignoring the graph structure (Moryossef et al., 2019; Castro Ferreira et al., 2019).”
AggGen: Ordering and Aggregating while Generating (2021, background): “For example, in (Moryossef et al., 2019b; Zhao et al., 2020) the sentence plan is created independently, incurring error propagation; Wiseman et al.”
Improving Quality and Efficiency in Plan-based Neural Data-to-text Generation (2019, methodology): “step systems described in Moryossef et al. (2019), which are state of the art.”
Encode, Tag, Realize: High-Precision Text Editing (2019, background): “While classic works on text generation make a distinction between planning and realization, end-to-end neural approaches typically ignore this distinction, with the exception of few works (Moryossef et al., 2019; Puduppully et al., 2018).”
Understanding Points of Correspondence between Sentences for Abstractive Summarization (2020, background): “These systems rely on common words to derive a connected graph from input sentences or subject-verb-object triples (Moryossef et al., 2019).”
Plan-then-Generate: Controlled Data-to-Text Generation via Planning (2021, methodology): “The ﬁrst type of models does not use pre-trained language models (PLMs), including GTR-LSTM (Trisedya et al., 2018), Transformer (Ferreira et al., 2019), Step-by-Step (Moryossef et al., 2019), and PLANENC (Zhao et al., 2020).”
Neural Data-to-Text Generation via Jointly Learning the Segmentation and Correspondence (2020, methodology): “Our model is also in spirit related to recent attempts at separating content planning and surface realization in neural data-to-text models (Zhao et al., 2018; Puduppully et al., 2019; Moryossef et al., 2019; Fer-reira et al., 2019).”
Revisiting Challenges in Data-to-Text Generation with Fact Grounding (2020, background): “Moryossef et al. (2019) demonstrates isolating the content planning from the joint end-to-end training and employing multiple valid content plans during testing.”
Neural data-to-text generation: A comparison between pipeline and end-to-end architectures (2019, methodology): “UPF-FORGe is a proper pipeline system with several sequential steps, Melbourne ﬁrst generates a delexicalized template to later realize the re-ferring expressions, and Moryossef et al. (2019) splits the process up into Planning, where ordering and structuring are merged, and Realization.”
A Token-level Reference-free Hallucination Detection Benchmark for Free-form Text Generation (2021, background): “, 2020b), and template/scaffold guided schema with explicit plans (Ma et al., 2019; Moryossef et al., 2019; Balakrishnan et al., 2019; Du et al., 2020; Liu et al., 2021), e.”
A Tree-to-Sequence Model for Neural NLG in Task-Oriented Dialog (2019, methodology): “Moryossef et al. (2019) add a symbolic text planning step where facts are grouped and ordered in the input; in contrast to our work though, their approach uses standard Seq2Seq models for realization and leaves no ordering choices to the model.”
Have Your Text and Use It Too! End-to-End Neural Data-to-Text Generation with Semantic Fidelity (2020, background): “Another approach has been to rely on heuristics that map data values to potential realizations in the text, thus computing a Slot Error Rate (SER) metric (Duˇsek et al., 2019; Juraska et al., 2019; Moryossef et al., 2019a).”