• Publications
  • Influence
Hypothesis Only Baselines in Natural Language Inference
TLDR
We propose a hypothesis only baseline for diagnosing Natural Language Inference (NLI). Expand
  • 219
  • 35
  • PDF
Gender Bias in Coreference Resolution
TLDR
We present a novel, Winograd schema-style set of minimal pair sentences that differ only by pronoun gender. Expand
  • 152
  • 24
  • PDF
On Measuring Social Biases in Sentence Encoders
TLDR
The Word Embedding Association Test shows that GloVe and word2vec word embeddings exhibit human-like implicit biases based on gender, race, and other social constructs. Expand
  • 96
  • 13
  • PDF
Collecting Diverse Natural Language Inference Problems for Sentence Representation Evaluation
TLDR
We present a large-scale collection of diverse natural language inference (NLI) datasets that help provide insight into how well a sentence representation captures distinct types of reasoning. Expand
  • 76
  • 13
  • PDF
Ordinal Common-sense Inference
TLDR
We propose an evaluation of automated common-sense inference based on an extension of recognizing textual entailment: predicting ordinal human responses on the subjective likelihood of an inference holding in a given context. Expand
  • 71
  • 11
  • PDF
Script Induction as Language Modeling
TLDR
We argue that the narrative cloze can be productively reframed as a language modeling task. Expand
  • 53
  • 11
  • PDF
Semantic Proto-Roles
TLDR
We present the first large-scale, corpus based verification of Dowty’s seminal theory of proto-roles. Expand
  • 56
  • 10
  • PDF
Universal Decompositional Semantics on Universal Dependencies
TLDR
This paper describes the Universal Decompositional Semantics (Decomp) project, which aims to augments Universal Dependencies (UD) data sets with robust, scalable semantic annotations based in linguistic theory. Expand
  • 82
  • 6
  • PDF
Neural models of factuality
TLDR
We present two neural models for event factuality prediction, which yield significant performance gains over previous models on three eventfactuality datasets: FactBank, UW, and MEANTIME. Expand
  • 36
  • 5
  • PDF
...
1
2
3
4
...