SOTAVerified

Data-to-Text Generation

A classic problem in natural-language generation (NLG) involves taking structured data, such as a table, as input, and producing text that adequately and fluently describes this data as output. Unlike machine translation, which aims for complete transduction of the sentence to be translated, this form of NLG is usually taken to require addressing (at least) two separate challenges: what to say, the selection of an appropriate subset of the input data to discuss, and how to say it, the surface realization of a generation.

( Image credit: Data-to-Text Generation with Content Selection and Planning )

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 51–100 of 219 papers

TitleStatusHype
Selective Token Generation for Few-shot Natural Language GenerationCode0
High Recall Data-to-text Generation with Progressive Edit—0
Innovations in Neural Data-to-text Generation: A Survey—0
Leveraging Natural Supervision for Language Representation Learning and GenerationCode1
Neural Data-to-Text Generation Based on Small Datasets: Comparing the Added Value of Two Semi-Supervised Learning Approaches on Top of a Large Language Model—0
Curriculum-Based Self-Training Makes Better Few-Shot Learners for Data-to-Text GenerationCode0
Generating Textual Explanations for Machine Learning Models Performance: A Table-to-Text Task—0
CoNT: Contrastive Neural Text GenerationCode2
R2D2: Robust Data-to-Text with Replacement DetectionCode0
What Makes Data-to-Text Generation Hard for Pretrained Language Models?—0
Robust (Controlled) Table-to-Text Generation with Structure-Aware Equivariance LearningCode1
How Do Seq2Seq Models Perform on End-to-End Data-to-Text Generation?Code0
uFACT: Unfaithful Alien-Corpora Training for Semantically Consistent Data-to-Text Generation—0
NMTScore: A Multilingual Analysis of Translation-based Text Similarity MeasuresCode1
GAP: A Graph-aware Language Model Framework for Knowledge Graph-to-Text GenerationCode1
Neural Pipeline for Zero-Shot Data-to-Text GenerationCode1
Chart-to-Text: A Large-Scale Benchmark for Chart SummarizationCode1
Faithfulness in Natural Language Generation: A Systematic Survey of Analysis, Evaluation and Optimization Methods—0
Data-to-text Generation with Variational Sequential PlanningCode1
Survey of Hallucination in Natural Language Generation—0
XAlign: Cross-lingual Fact-to-Text Alignment and Generation for Low-Resource LanguagesCode1
A Survey of Natural Language Generation—0
May the Force Be with Your Copy Mechanism: Enhanced Supervised-Copy Method for Natural Language Generation—0
Does Structure Matter? Leveraging Data-to-Text Generation for Answering Complex Information NeedsCode0
Search and Learn: Improving Semantic Coverage for Data-to-Text GenerationCode0
Cross-lingual Alignment of Knowledge Graph Triples with Sentences—0
Improving Compositional Generalization with Self-Training for Data-to-Text Generation—0
Neural Pipeline for Zero-Shot Data-to-Text Generation—0
How Do Seq2Seq Models Perform on End-to-End Data-to-Text Generation?—0
Open Domain Question Answering over Virtual Documents: A Unified Approach for Data and Text—0
Grouped-Attention for Content-Selection and Content-Plan Generation—0
TWT: Table with Written Text for Controlled Data-to-Text Generation—0
Entity-Based Semantic Adequacy for Data-to-Text Generation—0
Open Domain Question Answering with A Unified Knowledge InterfaceCode1
Improving Compositional Generalization with Self-Training for Data-to-Text GenerationCode0
Control Prefixes for Parameter-Efficient Text GenerationCode1
Selective Token Generation for Few-shot Language Modeling—0
Perturbation CheckLists for Evaluating NLG Evaluation MetricsCode0
Biomedical Data-to-Text Generation via Fine-Tuning TransformersCode1
Plan-then-Generate: Controlled Data-to-Text Generation via PlanningCode1
Evaluation Guidelines to Deal with Implicit Phenomena to Assess Factuality in Data-to-Text Generation—0
NUIG-DSI’s submission to The GEM Benchmark 2021—0
Exploring Structural Encoding for Data-to-Text Generation—0
Improving Encoder by Auxiliary Supervision Tasks for Table-to-Text GenerationCode1
HTLM: Hyper-Text Pre-Training and Prompting of Language Models—0
On Training Instance Selection for Few-Shot Neural Text Generation—0
Stage-wise Fine-tuning for Graph-to-Text GenerationCode1
R2D2: Relational Text Decoding with Transformers—0
Learning to Reason for Text Generation from Scientific TablesCode1
Data-QuestEval: A Referenceless Metric for Data-to-Text Semantic EvaluationCode1
Show:102550
← PrevPage 2 of 5Next →
All datasetsWebNLGE2E NLG ChallengeWebNLG FullCleaned E2E NLG ChallengeRotoWireRotoWire (Relation Generation)ToTToXAlignDARTMULTIWOZ 2.1RotoWire (Content Ordering)Rotowire (Content Selection)

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1Control Prefixes (A1, T5-large)BLEU67.32—Unverified
2Control Prefixes (A1, A2, T5-large)BLEU67.15—Unverified
3JointGT BaselineBLEU67.08—Unverified
4FactT5BBLEU67.04—Unverified
5T5B BaselineBLEU67.04—Unverified
6FactJointGTBLEU66.89—Unverified
7T5-large + Wiki + PositionBLEU66.07—Unverified
8HTML (fine-tuning)BLEU65.4—Unverified
9T5-smallBLEU65.05—Unverified
10TrICy (trK = trk* = 0.24)BLEU64.73—Unverified
#ModelMetricClaimedVerifiedStatus
1S_1^RBLEU68.6—Unverified
2EDA_CSBLEU67.05—Unverified
3TrICy (trK = 0)BLEU66.43—Unverified
4SlugBLEU66.19—Unverified
5TGenBLEU65.93—Unverified
6EDA_CS (TL)BLEU65.8—Unverified
7Sys1-PrimaryBLEU65.61—Unverified
8ZhangBLEU65.45—Unverified
9Self-memoryBLEU65.11—Unverified
10GongBLEU64.22—Unverified
#ModelMetricClaimedVerifiedStatus
1Control Prefixes (A1, A2, T5-large)BLEU62.27—Unverified
2Control Prefixes (A1, T5-large)BLEU61.94—Unverified
3T5-large + Wiki + PositionBLEU60.56—Unverified
4T5-largeBLEU59.7—Unverified
5T5-LargeBLEU57.1—Unverified
6HTLM (prefix 0.1%)BLEU56.3—Unverified
7DATATUNER_NO_FCBLEU52.9—Unverified
8Transformer (Pipeline)BLEU51.68—Unverified
#ModelMetricClaimedVerifiedStatus
1Control Prefixes (T5-large)BLEU (Test set)44.15—Unverified
2DataTuner_FCBLEU (Test set)43.6—Unverified
3TGenBLEU (Test set)40.73—Unverified
4LSTMMETEOR (Validation set)0.39—Unverified
5TGenMETEOR (Validation set)0.39—Unverified
6BARTMETEOR (Validation set)0.37—Unverified
7T5METEOR (Validation set)0.37—Unverified
#ModelMetricClaimedVerifiedStatus
1HierarchicalEncoder + NR + IRBLEU17.96—Unverified
2Hierarchical transformer encoder + conditional copyBLEU17.5—Unverified
3Force-CopyBLEU17.26—Unverified
4Neural Content Planning + conditional copyBLEU16.5—Unverified
5MacroBLEU15.46—Unverified
6Encoder-decoder + conditional copyBLEU14.19—Unverified
#ModelMetricClaimedVerifiedStatus
1SeqPlanPrecision97.6—Unverified
2MacroPrecision97.6—Unverified
3Force-CopyPrecision95.4—Unverified
4Hierarchical Transformer Encoder + conditional copyPrecision89.46—Unverified
5Neural Content Planning + conditional copyPrecision87.47—Unverified
6Encoder-decoder + conditional copyPrecision74.8—Unverified
#ModelMetricClaimedVerifiedStatus
1T5-3BBLEU49.5—Unverified
2LATTICE (T5-base)BLEU48.4—Unverified
3BERT-to-BERTBLEU44—Unverified
4Pointer GeneratorBLEU41.6—Unverified
5NCP+CC (Puduppully et al 2019)BLEU19.2—Unverified
6T5METEOR0.36—Unverified
#ModelMetricClaimedVerifiedStatus
1Fact-aware embedding with mT5BLEU429.27—Unverified
2Bi-lingual mT5BLEU425.88—Unverified
3mT5BLEU425—Unverified
4Vanilla TransformerBLEU419.9—Unverified
5Translate-Output mT5BLEU418.91—Unverified
6Graph Attention Network Encoder +Transformer DecoderBLEU418.3—Unverified
#ModelMetricClaimedVerifiedStatus
1T5B BaselineBLEU48.47—Unverified
2FactT5BBLEU48.37—Unverified
3self-mem + new dataBLEU47.76—Unverified
4JointGT BaselineBLEU47.51—Unverified
5FactJointGTBLEU47.39—Unverified
#ModelMetricClaimedVerifiedStatus
1T5-BaseBLEU35.1—Unverified
2T5-smallBLEU34.96—Unverified
3T2G2BLEU34.91—Unverified
4SC-GPT2BLEU30.76—Unverified
5HDSABLEU26.48—Unverified
#ModelMetricClaimedVerifiedStatus
1Hierarchical Transformer Encoder + conditional copyDLD18.9—Unverified
2Neural Content Planning + conditional copyDLD18.58—Unverified
3MacroDLD17.7—Unverified
4Force-CopyDLD17.26—Unverified
5Encoder-decoder + conditional copyDLD8.68—Unverified
#ModelMetricClaimedVerifiedStatus
1Hierarchical Transformer Encoder + conditional copyPrecision39.47—Unverified
2Force-CopyPrecision34.34—Unverified
3Neural Content Planning + conditional copyPrecision34.18—Unverified
4MacroPrecision34.1—Unverified
5Encoder-decoder + conditional copyPrecision29.49—Unverified
#ModelMetricClaimedVerifiedStatus
1SeqPlanBLEU14.29—Unverified
2MacroBLEU12.62—Unverified
3ENTBLEU11.5—Unverified
4Force-CopyBLEU10.5—Unverified
#ModelMetricClaimedVerifiedStatus
1SeqPlanDLD22.7—Unverified
2MacroDLD21.8—Unverified
3Force-CopyDLD21.16—Unverified
4ENTDLD20.7—Unverified
#ModelMetricClaimedVerifiedStatus
1SeqPlanPrecision95.9—Unverified
2MacroPrecision94.4—Unverified
3Force-CopyPrecision84.5—Unverified
4ENTPrecision81.1—Unverified
#ModelMetricClaimedVerifiedStatus
1binmtBLEU score26.35—Unverified
2tgenBLEU score21.96—Unverified
3massBLEU score17.72—Unverified
#ModelMetricClaimedVerifiedStatus
1Force-CopyPrecision49.39—Unverified
2SeqPlanPrecision43.3—Unverified
3MacroPrecision40.8—Unverified
#ModelMetricClaimedVerifiedStatus
1self-mem + new data (random)METEOR46.11—Unverified
2self-mem + new data (fixed)METEOR46.07—Unverified
#ModelMetricClaimedVerifiedStatus
1Transition based Deep Input LinearizationBLEU80.49—Unverified
2GCN + featBLEU0.67—Unverified
#ModelMetricClaimedVerifiedStatus
1DataTuner_FCBLEU53.6—Unverified
2Bo3BLEU52.1—Unverified
#ModelMetricClaimedVerifiedStatus
1mBARTMETEOR0.46—Unverified
2mT5METEOR0.29—Unverified
#ModelMetricClaimedVerifiedStatus
1mBARTMETEOR0.61—Unverified
2mT5METEOR0.18—Unverified
#ModelMetricClaimedVerifiedStatus
1StructAdaptBleu48—Unverified
#ModelMetricClaimedVerifiedStatus
1T5-largeBLEU45.85—Unverified
#ModelMetricClaimedVerifiedStatus
1T5-largeBLEU69.27—Unverified
#ModelMetricClaimedVerifiedStatus
1OursBLEU24.56—Unverified