AllenNLP: A Deep Semantic Natural Language Processing Platform

By Matt Gardner, Joel Grus, Mark Neumann, Oyvind Tafjord, Pradeep Dasigi, Nelson F. Liu, Matthew E. Peters, Michael Schmitz, Luke S. Zettlemoyer. 2018

1,333 citations recorded by the source corpus.

Modern natural language processing (NLP) research requires writing code. Ideally this code would provide a precise definition of the approach, easy repeatability of results, and a basis for extending the research. However, many research codebases bury high-level parameters under implementation details, are challenging to run and debug, and are difficult enough to extend that they are more likely to be rewritten. This paper describes AllenNLP, a library for applying deep learning methods to NLP research that addresses these issues with easy-to-use command-line tools, declarative configuration-driven experiments, and modular NLP abstractions. AllenNLP has already increased the rate of research experimentation and the sharing of NLP components at the Allen Institute for Artificial Intelligence, and we are working to have the same impact across the field.

Subfields: Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Natural Language Processing

Citations in the corpus

Dissecting Contextual Word Embeddings: Architecture and Representation (2018, methodology): “Semantic Role Labeling The SRL model uses the reimplementation of He et al. (2017) from Gardner et al. (2018).”
SWAG: A Large-Scale Adversarial Dataset for Grounded Commonsense Inference (2018, methodology): “We implemented the neural models in PyTorch us-ing the AllenNLP library (Gardner et al., 2018”
GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding (2018, methodology): “We implement our models in the AllenNLP library (Gardner et al., 2017).”
Pretrained Transformers Improve Out-of-Distribution Robustness (2020, methodology): “We implement our models in AllenNLP (Gardner et al., 2018) and tune the hyperparameters to maximize validation performance on the IID task.”
SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems (2019, methodology): “Software Tools To facilitate using SuperGLUE, we release jiant (Wang et al., 2019b), 4 a modular software toolkit, built with PyTorch (Paszke et al., 2017), components from AllenNLP (Gardner et al., 2017), and the transformers package.”
DROP: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphs (2019, methodology): “We used the AllenNLP implementations of state-of-theart models for all of these representations (Gardner et al., 2017; Dozat et al., 2017; He et al., 2017; Stanovsky et al., 2018).”
Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inference (2019, methodology): “Implementation and evaluation: For DA and ESIM, we used the implementations from Al-lenNLP (Gardner et al., 2017).”
ReCoRD: Bridging the Gap between Human and Machine Commonsense Reading Comprehension (2018, methodology): “We then ran the state-of-the-art named entity recognition model (Peters et al., 2018) and the end-to-end coreference resolution model (Lee et al., 2017) provided by AllenNLP (Gardner et al., 2018) to annotate the crawled articles.”
QuAC: Question Answering in Context (2018, methodology): “The AllenNLP (Gardner et al., 2017) implementation we use reaches 82.”
Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answering (2018, methodology): “Our neural models are implemented with Al-lenNLP 15 (Gardner et al., 2017) and PyTorch 16 (Paszke et al., 2017).”
To Tune or Not to Tune? Adapting Pretrained Representations to Diverse Tasks (2019, methodology): “SA We used the implementation of the bi-attentive classiﬁcation network in AllenNLP (Gardner et al., 2017) with default hyper-parameters, except for tuning those noted above.”
Can You Tell Me How to Get Past Sesame Street? Sentence-Level Pretraining Beyond Language Modeling (2018, methodology): “We implement our models using the jiant toolkit, 3 which is in turn built on AllenNLP (Gard-ner et al., 2017) and on a public PyTorch implementation of BERT.”
Harms of Gender Exclusivity and Challenges in Non-Binary Representation in Language Technologies (2021, positive): “We provided survey participants with the AllenNLP Named Entity Recognition demo (Gardner et al., 2018).”
Toward Gender-Inclusive Coreference Resolution (2019, background): “Gender is also an object of study in gender recognition systems [Hamidi et al. 2018, e.g.,]. Much of this work has focused on gender bias with a (usually implicit) binary lens, an issue which was also called out recently by Larson [2017]. Outside of NLP, there have been many studies looking at how g”
HuggingFace's Transformers: State-of-the-art Natural Language Processing (2019, neutral): “At the time of writing, the authors have been mostly aware of FastBert 9 , FARM 10 , flair (Akbik et al., 2018, 2019), AllenNLP (Gardner et al., 2018) and PyText 11 but there are likely more interesting developments to be found, from research and internal projects to production packages.”
Finding Generalizable Evidence by Learning to Convince Q&A Models (2019, methodology): “We build off AllenNLP (Gardner et al., 2018) using PyTorch (Paszke et al.”