# Generalizing Natural Language Analysis through Span-relation Representations

arXiv: https://arxiv.org/abs/1911.03822 · HTML: https://mlsanity.com/paper/1911.03822/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 10 Nov 2019.

## Built on by (most cited first, 22 of 22)

- [PAQ: 65 Million Probably-Asked Questions and What You Can Do With Them](https://mlsanity.com/paper/2102.07033/) (2021)
- [Document-level Relation Extraction as Semantic Segmentation](https://mlsanity.com/paper/2106.03618/) (2021)
- [SpanNER: Named Entity Re-/Recognition as Span Prediction](https://mlsanity.com/paper/2106.00641/) (2021)
- [Packed Levitated Marker for Entity and Relation Extraction](https://mlsanity.com/paper/2109.06067/) (2021)
- [Weakly Supervised Named Entity Tagging with Learnable Logical Rules](https://mlsanity.com/paper/2107.02282/) (2021)
- [To Share or not to Share: Predicting Sets of Sources for Model Transfer Learning](https://mlsanity.com/paper/2104.08078/) (2021)
- [Improving Self-training for Cross-lingual Named Entity Recognition with Contrastive and Prototype Learning](https://mlsanity.com/paper/2305.13628/) (2023)
- [Preserving Knowledge Invariance: Rethinking Robustness Evaluation of Open Information Extraction](https://mlsanity.com/paper/2305.13981/) (2023)
- [Split-NER: Named Entity Recognition via Two Question-Answering-based Classifications](https://mlsanity.com/paper/2310.19942/) (2023)
- [A Data-Centric Framework for Composable NLP Workflows](https://mlsanity.com/paper/2103.01834/) (2021)
- [A Generative Model for Relation Extraction and Classification](https://mlsanity.com/paper/2202.13229/) (2022)
- [Span Classification with Structured Information for Disfluency Detection in Spoken Utterances](https://mlsanity.com/paper/2203.16028/) (2022)
- [Larger-Context Tagging: When and Why Does It Work?](https://mlsanity.com/paper/2104.04434/) (2021)
- [DualNER: A Dual-Teaching framework for Zero-shot Cross-lingual Named Entity Recognition](https://mlsanity.com/paper/2211.08104/) (2022)
- [ELIT: Emory Language and Information Toolkit](https://mlsanity.com/paper/2109.03903/) (2021)
- [Distilling Task-specific Logical Rules from Large Pre-trained Models](https://mlsanity.com/paper/2210.02768/) (2022)
- [Fast Rule-Based Decoding: Revisiting Syntactic Rules in Neural Constituency Parsing](https://mlsanity.com/paper/2212.08458/) (2022)
- [CLaC at SemEval-2023 Task 2: Comparing Span-Prediction and Sequence-Labeling approaches for NER](https://mlsanity.com/paper/2305.03845/) (2023)
- [Comparing and combining some popular NER approaches on Biomedical tasks](https://mlsanity.com/paper/2305.19120/) (2023)
- [Unifying Token and Span Level Supervisions for Few-Shot Sequence Labeling](https://mlsanity.com/paper/2307.07946/) (2023)
- [DiMB-RE: Mining the Scientific Literature for Diet-Microbiome Associations](https://mlsanity.com/paper/2409.19581/) (2024)
- [Nested Named Entity Recognition from Medical Texts: An Adaptive Shared Network Architecture with Attentive CRF](https://mlsanity.com/paper/2211.04759/) (2022)

## Abstract (arXiv)

Natural language processing covers a wide variety of tasks predicting syntax, semantics, and information content, and usually each type of output is generated with specially designed architectures. In this paper, we provide the simple insight that a great variety of tasks can be represented in a single unified format consisting of labeling spans and relations between spans, thus a single task-independent model can be used across different tasks. We perform extensive experiments to test this insight on 10 disparate tasks spanning dependency parsing (syntax), semantic role labeling (semantics), relation extraction (information content), aspect based sentiment analysis (sentiment), and many others, achieving performance comparable to state-of-the-art specialized models. We further demonstrate benefits of multi-task learning, and also show that the proposed method makes it easy to analyze differences and similarities in how the model handles different tasks. Finally, we convert these datasets into a unified format to build a benchmark, which provides a holistic testbed for evaluating future models for generalized natural language analysis.