# A Modest Pareto Optimisation Analysis of Dependency Parsers in 2021

arXiv: https://arxiv.org/abs/2106.04216 · HTML: https://mlsanity.com/paper/2106.04216/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 21 of its 64 references that are in the map.
- First posted on arXiv 8 Jun 2021.

## Builds on (most cited first, 21 of 21)

- [Energy and Policy Considerations for Deep Learning in NLP](https://mlsanity.com/paper/1906.02243/) (2019)
- [Deep Biaffine Attention for Neural Dependency Parsing](https://mlsanity.com/paper/1611.01734/) (2016)
- [Learning Word Vectors for 157 Languages](https://mlsanity.com/paper/1802.06893/) (2018)
- [Semi-Supervised Sequence Modeling with Cross-View Training](https://mlsanity.com/paper/1809.08370/) (2018)
- [Simple and Accurate Dependency Parsing Using Bidirectional LSTM Feature Representations](https://mlsanity.com/paper/1603.04351/) (2016)
- [Analogical Reasoning on Chinese Morphological and Semantic Relations](https://mlsanity.com/paper/1805.06504/) (2018)
- [Efficient Second-Order TreeCRF for Neural Dependency Parsing](https://mlsanity.com/paper/2005.00975/) (2020)
- [Head-Driven Phrase Structure Grammar Parsing on Penn Treebank](https://mlsanity.com/paper/1907.02684/) (2019)
- [Stack-Pointer Networks for Dependency Parsing](https://mlsanity.com/paper/1805.01087/) (2018)
- [Viable Dependency Parsing as Sequence Labeling](https://mlsanity.com/paper/1902.10505/) (2019)
- [Left-to-Right Dependency Parsing with Pointer Networks](https://mlsanity.com/paper/1903.08445/) (2019)
- [HULK: An Energy Efficiency Benchmark Platform for Responsible Natural Language Processing](https://mlsanity.com/paper/2002.05829/) (2020)
- [Bracketing Encodings for 2-Planar Dependency Parsing](https://mlsanity.com/paper/2011.00596/) (2020)
- [Distilling an Ensemble of Greedy Dependency Parsers into One MST Parser](https://mlsanity.com/paper/1609.07561/) (2016)
- [82 Treebanks, 34 Models: Universal Dependency Parsing with Multi-Treebank Models](https://mlsanity.com/paper/1809.02237/) (2018)
- [An Investigation of the Interactions Between Pre-Trained Word Embeddings, Character Models and POS Tags in Dependency Parsing](https://mlsanity.com/paper/1808.09060/) (2018)
- [Is POS Tagging Necessary or Even Helpful for Neural Dependency Parsing?](https://mlsanity.com/paper/2003.03204/) (2020)
- [Efficient EUD Parsing](https://mlsanity.com/paper/2006.00838/) (2020)
- [Distilling Neural Networks for Greener and Faster Dependency Parsing](https://mlsanity.com/paper/2006.00844/) (2020)
- [On the Frailty of Universal POS Tags for Neural UD Parsers](https://mlsanity.com/paper/2010.01830/) (2020)
- [Optimal Size-Performance Tradeoffs: Weighing PoS Tagger Models](https://mlsanity.com/paper/2104.07951/) (2021)

## Built on by (most cited first, 9 of 9)

- [Towards Climate Awareness in NLP Research](https://mlsanity.com/paper/2205.05071/) (2022)
- [Stochastic Multiple Target Sampling Gradient Descent](https://mlsanity.com/paper/2206.01934/) (2022)
- [4 and 7-bit Labeling for Projective and Non-Projective Dependency Trees](https://mlsanity.com/paper/2310.14319/) (2023)
- [On the Challenges of Fully Incremental Neural Dependency Parsing](https://mlsanity.com/paper/2309.16254/) (2023)
- [A Syntax-Injected Approach for Faster and More Accurate Sentiment Analysis](https://mlsanity.com/paper/2406.15163/) (2024)
- [Dancing in the syntax forest: fast, accurate and explainable sentiment analysis with SALSA](https://mlsanity.com/paper/2406.16071/) (2024)
- [Splitting EUD graphs into trees: A quick and clatty approach](https://mlsanity.com/paper/2106.13155/) (2021)
- [Parsing linearizations appreciate PoS tags - but some are fussy about errors](https://mlsanity.com/paper/2210.15219/) (2022)
- [The Fragility of Multi-Treebank Parsing Evaluation](https://mlsanity.com/paper/2209.06699/) (2022)

## Abstract (arXiv)

We evaluate three leading dependency parser systems from different paradigms on a small yet diverse subset of languages in terms of their accuracy-efficiency Pareto front. As we are interested in efficiency, we evaluate core parsers without pretrained language models (as these are typically huge networks and would constitute most of the compute time) or other augmentations that can be transversally applied to any of them. Biaffine parsing emerges as a well-balanced default choice, with sequence-labelling parsing being preferable if inference speed (but not training energy cost) is the priority.