A Survey of Active Learning for Natural Language Processing

2022/10/18 by Zhisong Zhang, Emma Strubell, Zhang, Zhisong +3 · 5 citations
Computer Science · #Machine Learning and Algorithms #Natural Language Processing Techniques #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2210.10109

Abstract

In this work, we provide a survey of active learning (AL) for its applications in natural language processing (NLP). In addition to a fine-grained categorization of query strategies, we also investigate several other important aspects of applying AL to NLP problems. These include AL for structured prediction tasks, annotation cost, model learning (especially with deep neural models), and starting and stopping AL. Finally, we conclude with a discussion of related topics and future directions.

Cited by

• Topic-Specific Classifiers are Better Relevance Judges than Prompted LLMs
• A Survey of the State-of-the-Art in Conversational Question Answering Systems
• PromptAL: Sample-Aware Dynamic Soft Prompts for Few-Shot Active Learning
• ART: Adaptive Relation Tuning for Generalized Relation Prediction
• Actively evaluating and learning the distinctions that matter: Vaccine safety signal detection from emergency triage notes

Related

• A Survey on Deep Active Learning: Recent Advances and New Frontiers
• Active Learning for Natural Language Generation
• Reassessing Active Learning Adoption in Contemporary NLP: A Community Survey
• Active Learning for New Domains in Natural Language Understanding
• From Selection to Generation: A Survey of LLM-based Active Learning