• Publications
  • Influence
Don't Stop Pretraining: Adapt Language Models to Domains and Tasks
TLDR
We present a study across four domains (biomedical and computer science publications, news, and reviews) and eight classification tasks, showing that a second phase of pretraining in-domain (domain-adaptive pretraining) leads to performance gains, under both high- and low-resource settings. Expand
  • 134
  • 27
  • PDF
Quoref: A Reading Comprehension Dataset with Questions Requiring Coreferential Reasoning
TLDR
We present a new crowdsourced dataset containing more than 24K span-selection questions that require resolving coreference among entities in over 4.7K English paragraphs from Wikipedia. Expand
  • 41
  • 10
  • PDF
A Mention-Ranking Model for Abstract Anaphora Resolution
TLDR
We propose a mention-ranking model that learns how abstract anaphors relate to their antecedents with an LSTM-Siamese Net. Expand
  • 16
  • 4
  • PDF
SRL4ORL: Improving Opinion Role Labelling using Multi-task Learning with Semantic Role Labeling
TLDR
We investigate the limitations of neural models in solving the fine-grained opinion analysis on MPQA and address this issue using different multi-task learning techniques. Expand
  • 13
  • 3
  • PDF
Explaining NLP Models via Minimal Contrastive Editing (MiCE)
TLDR
We present MINIMAL CONTRASTIVE EDITING (MICE), a method for generating contrastive explanations of model predictions in the form of edits to inputs that change model outputs to the contrast case. Expand
  • 1
  • 1
  • PDF
Multilingual Modal Sense Classification using a Convolutional Neural Network
TLDR
We explore a CNN architecture for classifying modal sense in English and German. Expand
  • 15
  • PDF
Modal Sense Classification At Large: Paraphrase-Driven Sense Projection, Semantically Enriched Classification Models and Cross-Genre Evaluations
TLDR
We build a semantically enriched model for modal sense classification by designing novel features related to lexical, proposition-level and discourse-level semantic factors. Expand
  • 8
Measuring Association Between Labels and Free-Text Rationales
TLDR
We propose measurements of label-rationale association, a necessary property of faithful rationales, for self-rationalizing models that provide free-text natural language rationales. Expand
  • 2
  • PDF
Natural Language Rationales with Full-Stack Visual Reasoning: From Pixels to Semantic Frames to Commonsense Graphs
TLDR
We present Rationale^VT Transformer, an integrated model that learns to generate free-text rationales by combining pretrained language models with object recognition, grounded visual semantic frames, and visual commonsense graphs. Expand
  • 1
  • PDF
Easy, Reproducible and Quality-Controlled Data Collection with Crowdaq
TLDR
We introduce CROWDAQ, an open-source platform that standardizes the data collection pipeline with customizable user-interface components, automated annotator qualification, and saved pipelines in a re-usable format. Expand
  • 1
  • PDF