TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/BiLSTM

BiLSTM

Reported on 30 benchmarks across 11 tasks · 7 papers · 4 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing19 results

  • Relation ExtractiononDocRED
    F1· 2019-06-14
    51.06
    best: 67.53 (DREEAM)
    SOTA
    DocRED: A Large-Scale Document-Level Relation Extraction DatasetarXiv:1906.06127
  • Relation ExtractiononDocRED
    Ign F1· 2019-06-14
    44.73
    best: 65.47 (DREEAM)
    SOTA
    DocRED: A Large-Scale Document-Level Relation Extraction DatasetarXiv:1906.06127
  • Poem meters classificationonPCD
    Accuracy· 2019-05-07
    96.38
    SOTA
    Learning meters of Arabic and English poems with Recurrent Neural Networks: a step forward for language understanding and synthesisarXiv:1905.05700
  • Text ClassificationonCivil Comments
    GMB Subgroup· 2023-01-26
    0.8636
    best: 0.8807 (RoBERTa Focal Loss)
    A benchmark for toxic comment classification on Civil Comments datasetarXiv:2301.11125
  • Text ClassificationonCivil Comments
    Micro F1· 2023-01-26
    0.5115
    best: 0.5958 (Unfreeze Glove ResNet 44)
    A benchmark for toxic comment classification on Civil Comments datasetarXiv:2301.11125
  • Text ClassificationonCivil Comments
    Precision· 2023-01-26
    0.3572
    best: 0.4835 (Unfreeze Glove ResNet 44)
    A benchmark for toxic comment classification on Civil Comments datasetarXiv:2301.11125
  • Relation ExtractiononDialogRE
    F1 (v1)· 2020-04-17
    48.6
    best: 75.1 (GRASP_Large)
    Dialogue-Based Relation ExtractionarXiv:2004.08056
  • Relation ExtractiononDialogRE
    F1c (v1)· 2020-04-17
    45
    best: 66.7 (GRASP_Large)
    Dialogue-Based Relation ExtractionarXiv:2004.08056
  • Relation ExtractiononDocRED
    F1· 2019-06-14
    50.12
    best: 67.53 (DREEAM)
    DocRED: A Large-Scale Document-Level Relation Extraction DatasetarXiv:1906.06127
  • Relation ExtractiononDocRED
    Ign F1· 2019-06-14
    43.6
    best: 65.47 (DREEAM)
    DocRED: A Large-Scale Document-Level Relation Extraction DatasetarXiv:1906.06127
  • Machine TranslationonWMT2016 English-Romanian
    BLEU score· 2016-11-07
    27.5
    best: 34.7 (DeLighT)
    A Convolutional Encoder Model for Neural Machine TranslationarXiv:1611.02344
  • Sentence EmbeddingsonGoogle Dataset
    CR
    0.43
  • Sentence EmbeddingsonGoogle Dataset
    F1
    0.8
    best: 0.855 (SLAHAN (LSTM+syntactic-information))
  • Text ClassificationonMVICTOR (type)
    Average F1
    0.7092
    best: 0.7505 (CNN + CRF)
  • Text ClassificationonMVICTOR (type)
    Weighted F1
    0.9433
    best: 0.9537 (CNN + CRF)
  • Text ClassificationonSVICTOR (type)
    Average F1
    0.7281
    best: 0.774 (CNN + CRF)
  • Text ClassificationonSVICTOR (type)
    Weighted F1
    0.9465
    best: 0.9533 (CNN + CRF)
  • Sentence CompressiononGoogle Dataset
    CR
    0.43
  • Sentence CompressiononGoogle Dataset
    F1
    0.8
    best: 0.855 (SLAHAN (LSTM+syntactic-information))

Methodology9 results

  • ClassificationonCivil Comments
    GMB Subgroup· 2023-01-26
    0.8636
    best: 0.8807 (RoBERTa Focal Loss)
    A benchmark for toxic comment classification on Civil Comments datasetarXiv:2301.11125
  • ClassificationonCivil Comments
    Micro F1· 2023-01-26
    0.5115
    best: 0.5958 (Unfreeze Glove ResNet 44)
    A benchmark for toxic comment classification on Civil Comments datasetarXiv:2301.11125
  • ClassificationonCivil Comments
    Precision· 2023-01-26
    0.3572
    best: 0.4835 (Unfreeze Glove ResNet 44)
    A benchmark for toxic comment classification on Civil Comments datasetarXiv:2301.11125
  • Representation LearningonGoogle Dataset
    CR
    0.43
  • Representation LearningonGoogle Dataset
    F1
    0.8
    best: 0.855 (SLAHAN (LSTM+syntactic-information))
  • ClassificationonMVICTOR (type)
    Average F1
    0.7092
    best: 0.7505 (CNN + CRF)
  • ClassificationonMVICTOR (type)
    Weighted F1
    0.9433
    best: 0.9537 (CNN + CRF)
  • ClassificationonSVICTOR (type)
    Average F1
    0.7281
    best: 0.774 (CNN + CRF)
  • ClassificationonSVICTOR (type)
    Weighted F1
    0.9465
    best: 0.9533 (CNN + CRF)

Knowledge Base2 results

  • Text SummarizationonGoogle Dataset
    CR
    0.43
  • Text SummarizationonGoogle Dataset
    F1
    0.8
    best: 0.855 (SLAHAN (LSTM+syntactic-information))

Medical1 result

  • Drug Discoveryonclintox
    AUC· 2024-07-08
    0.97
    SOTA
    Accelerating Drug Safety Assessment using Bidirectional-LSTM for SMILES DataarXiv:2407.18919

Audio1 result

  • Text-To-Speech SynthesisonHelsinki Prosody Corpus
    Accuracy· 2019-08-06
    82.1
    best: 83.2 (BERT)
    Predicting Prosodic Prominence from Text with Pre-trained Contextualized Word RepresentationsarXiv:1908.02262