mlsanityData to 7 Oct 2026

Universal Natural Language Processing with Limited Annotations: Try Few-shot Textual Entailment as a Start

arXiv 2010.02584 ↗First posted 6 Oct 2020cs.CLcited by 44

Not on the map. Why?

Abstract (arXiv)

A standard way to address different NLP problems is by first constructing a problem-specific dataset, then building a model to fit this dataset. To build the ultimate artificial intelligence, we desire a single machine that can handle diverse new problems, for which task-specific annotations are limited. We bring up textual entailment as a unified solver for such NLP problems. However, current research of textual entailment has not spilled much ink on the following questions: (i) How well does a pretrained textual entailment system generalize across domains with only a handful of domain-specific examples? and (ii) When is it worth transforming an NLP task into textual entailment? We argue that the transforming is unnecessary if we can obtain rich annotations for this task. Textual entailment really matters particularly when the target NLP task has insufficient annotations. Universal NLP can be probably achieved through different routines. In this work, we introduce Universal Few-shot textual Entailment (UFO-Entail). We demonstrate that this framework enables a pretrained entailment model to work well on new entailment domains in a few-shot setting, and show its effectiveness as a unified solver for several downstream NLP tasks such as question answering and coreference resolution when the end-task annotations are limited. Code: https://github.com/salesforce/UniversalFewShotNLP

Key facts. Cited by 44 papers in the map (data through 7 Oct 2026). First posted on arXiv 6 Oct 2020.

Builds on

No reference list yet.

Built on by

Cited by year

Citing papers per year: 2021: 13; 2022: 26; 2023: 5.

44 papers.

Sort:

By area:

  1. 2022Discovering Latent Knowledge in Language Models Without Supervisioncited by 733
  2. 2021Parameter-efficient Multi-task Fine-tuning for Transformers via Shared Hypernetworkscited by 278
  3. 2022UnifiedSKG: Unifying and Multi-Tasking Structured Knowledge Grounding with Text-to-Text Language Modelscited by 231
  4. 2022Generating Training Data with Language Models: Towards Zero-Shot Language Understandingcited by 199
  5. 2021CrossFit: A Few-shot Learning Challenge for Cross-task Generalization in NLPcited by 142
  6. 2021Entailment as Few-Shot Learnercited by 122
  7. 2021Differentiable Prompt Makes Pre-trained Language Models Better Few-shot Learnerscited by 115
  8. 2022State-of-the-art generalisation research in NLP: A taxonomy and reviewcited by 97
  9. 2021Label Verbalization and Entailment for Effective Zero- and Few-Shot Relation Extractioncited by 73
  10. 2022Improving In-Context Few-Shot Learning via Self-Supervised Trainingcited by 56
  11. 2021Discrete and Soft Prompting for Multilingual Modelscited by 49
  12. 2021Can NLI Models Verify QA Systems' Predictions?cited by 46
32 more
  1. 2021Paradigm Shift in Natural Language Processingcited by 39
  2. 2023Defending Pre-trained Language Models as Few-shot Learners against Backdoor Attackscited by 32
  3. 2022Summarization as Indirect Supervision for Relation Extractioncited by 31
  4. 2021Incremental Few-shot Text Classification with Multi-round New Classes: Formulation, Dataset and Systemcited by 30
  5. 2022Textual Entailment for Event Argument Extraction: Zero- and Few-Shot with Multi-Source Learningcited by 25
  6. 2022Ontology-enhanced Prompt-tuning for Few-shot Learningcited by 24
  7. 2023Building Efficient Universal Classifiers with Natural Language Inferencecited by 19
  8. 2022Ultra-fine Entity Typing with Indirect Supervision from Natural Language Inferencecited by 17
  9. 2022Few-Shot Learning with Siamese Networks and Label Tuningcited by 15
  10. 2022PART: Pre-trained Authorship Representation Transformercited by 13
  11. 2022Can NLI Provide Proper Indirect Supervision for Low-resource Biomedical Relation Extraction?cited by 13
  12. 2022Active Few-Shot Learning with FASLcited by 11
  13. 2021Diverse Distributions of Self-Supervised Tasks for Meta-Learning in NLPcited by 10
  14. 2021CLUES: Few-Shot Learning Evaluation in Natural Language Understandingcited by 9
  15. 2022Few-shot Multimodal Sentiment Analysis based on Multimodal Probabilistic Fusion Promptscited by 9
  16. 2023What do Language Models know about word senses? Zero-Shot WSD with Language Models and Domain Inventoriescited by 9
  17. 2021LMTurk: Few-Shot Learners as Crowdsourcing Workers in a Language-Model-as-a-Service Frameworkcited by 8
  18. 2023Adversarial Robustness of Prompt-based Few-Shot Learning for Natural Language Understandingcited by 8
  19. 2023Text Alignment Is An Efficient Unified Model for Massive NLP Taskscited by 8
  20. 2022OpenStance: Real-world Zero-shot Stance Detectioncited by 7
  21. 2022LIME: Weakly-Supervised Text Classification Without Seedscited by 6
  22. 2022Zero-Shot Aspect-Based Sentiment Analysiscited by 5
  23. 2022Template-based Approach to Zero-shot Intent Recognitioncited by 5
  24. 2022Zero and Few-shot Learning for Author Profilingcited by 4
  25. 2022Learning to Select from Multiple Optionscited by 4
  26. 2022Unsupervised Ranking and Aggregation of Label Descriptions for Zero-Shot Classifierscited by 1
  27. 2022Zero-shot Aspect-level Sentiment Classification via Explicit Utilization of Aspect-to-Document Sentiment Compositioncited by 1
  28. 2021Exploring the Limits of Natural Language Inference Based Setup for Few-Shot Intent Detectioncited by 0
  29. 2022To Tune or Not To Tune? Zero-shot Models for Legal Case Entailmentcited by 0
  30. 2022Revising Image-Text Retrieval via Multi-Modal Entailmentcited by 0
  31. 2022ConEntail: An Entailment-based Framework for Universal Zero and Few Shot Classification with Supervised Contrastive Pretrainingcited by 0
  32. 2022BERT-Flow-VAE: A Weakly-supervised Model for Multi-Label Text Classificationcited by 0