Data-to-text Generation with Macro Planning

Puduppully, Ratish; Lapata, Mirella

doi:10.1162/tacl_a_00381

Cited by 37 publications

(31 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As the improvement, neural two-stage models Puduppully et al, 2019;Moryossef et al, 2019;Puduppully and Lapata, 2021;Su et al, 2021) decompose the table-to-text generation into content planning and surface generation stages. In general, content planning is implemented by Pointer Networks (Vinyals et al, 2015).…”

Section: Related Workmentioning

confidence: 99%

Plan-then-Seam: Towards Efficient Table-to-Text Generation

Li¹,

Geng²,

Fang³

et al. 2023

Preprint

View full text Add to dashboard Cite

Table -to-text generation aims at automatically generating text to help people conveniently obtain salient information in tables. Recent works explicitly decompose the generation process into content planning and surface generation stages, employing two autoregressive networks for them respectively. However, they are computationally expensive due to the nonparallelizable nature of autoregressive decoding and the redundant parameters of two networks. In this paper, we propose the first totally non-autoregressive table-to-text model (Plan-then-Seam, PTS) that produces its outputs in parallel with one single network. PTS firstly writes and calibrates one plan of the content to be generated with a novel rethinking pointer predictor, and then takes the plan as the context for seaming to decode the description. These two steps share parameters and perform iteratively to capture token interdependency while keeping parallel decoding. Experiments on two public benchmarks show that PTS achieves 3.0 ∼ 5.6 times speedup for inference time, reducing 50% parameters, while maintaining as least comparable performance against strong two-stage table-to-text competitors 1 .

show abstract

Section: Related Workmentioning

confidence: 99%

Plan-then-Seam: Towards Efficient Table-to-Text Generation

Li¹,

Geng²,

Fang³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…We conduct experiments using five representative Seq2Seq models on four commonly used data-totext datasets and evaluate the generated texts accordingly 2 . Note that we do not use models that are designed for specific data sets or data structures (Moryossef et al, 2019;Rebuffel et al, 2020;Puduppully and Lapata, 2021), but adopt models that allow inputs of different formats and structures, which brings convenience to comparison on different data sets. Besides, most specific models for data-to-text generation are actually based on these typical Seq2Seq models (Ferreira et al, 2019;Rebuffel et al, 2020), which also proves the rationality of our selection of these models.…”

Section: Models and Datasetsmentioning

confidence: 99%

How Do Seq2Seq Models Perform on End-to-End Data-to-Text Generation?

Xunjian¹,

Wan²

2022

Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

View full text Add to dashboard Cite

With the rapid development of deep learning, Seq2Seq paradigm has become prevalent for end-to-end data-to-text generation, and the BLEU scores have been increasing in recent years. However, it is widely recognized that there is still a gap between the quality of the texts generated by models and the texts written by human. In order to better understand the ability of Seq2Seq models, evaluate their performance and analyze the results, we choose to use Multidimensional Quality Metric(MQM) to evaluate several representative Seq2Seq models on end-to-end data-to-text generation. We annotate the outputs of five models on four datasets with eight error types and find that 1) copy mechanism is helpful for the improvement in Omission and Inaccuracy Extrinsic errors but it increases other types of errors such as Addition; 2) pre-training techniques are highly effective, and pre-training strategy and model size are very significant; 3) the structure of the dataset also influences the model's performance greatly; 4) some specific types of errors are generally challenging for seq2seq models.

show abstract

“…To alleviate this problem, one research direction adopts coarse-to-fine progressive text generation (Tan et al, 2021). This generation paradigm has been studied in many text generation systems for specific tasks, such as datato-text generation (Moryossef et al, 2019;Puduppully and Lapata, 2021), storytelling (Goldfarb-Tarrant et al, 2020;Orbach and Goldberg, 2020), and dialogue generation (Xu et al, 2020a). Our work adopts a generative event transition planner that is trained on a large amount of event transition paths, aiming to arrange the ensuing events in open-ended text generation.…”

Section: Related Workmentioning

confidence: 99%

Event Transition Planning for Open-ended Text Generation

Li¹,

Li²,

Bi³

et al. 2022

Findings of the Association for Computational Linguistics: ACL 2022

View full text Add to dashboard Cite

Open-ended text generation tasks, such as dialogue generation and story completion, require models to generate a coherent continuation given limited preceding context. The openended nature of these tasks brings new challenges to the neural auto-regressive text generators nowadays. Despite these neural models are good at producing human-like text, it is difficult for them to arrange causalities and relations between given facts and possible ensuing events. To bridge this gap, we propose a novel two-stage method which explicitly arranges the ensuing events in open-ended text generation. Our approach can be understood as a specially-trained coarse-to-fine algorithm, where an event transition planner provides a "coarse" plot skeleton and a text generator in the second stage refines the skeleton. Experiments on two open-ended text generation tasks demonstrate that our proposed method effectively improves the quality of the generated text, especially in coherence and diversity. The code is available at: https://github.com/ qtli/EventPlanforTextGen.

show abstract

Data-to-text Generation with Macro Planning

Cited by 37 publications

References 39 publications

Plan-then-Seam: Towards Efficient Table-to-Text Generation

Plan-then-Seam: Towards Efficient Table-to-Text Generation

How Do Seq2Seq Models Perform on End-to-End Data-to-Text Generation?

Event Transition Planning for Open-ended Text Generation

Contact Info

Product

Resources

About