• Publications
  • Influence
Abductive Commonsense Reasoning
TLDR
This study introduces a challenge dataset, ART, that consists of over 20k commonsense narrative contexts and 200k explanations, and conceptualizes two new tasks -- Abductive NLI: a multiple-choice question answering task for choosing the more likely explanation, and Abduction NLG: a conditional generation task for explaining given observations in natural language.
Blockwise Self-Attention for Long Document Understanding
TLDR
This model extends BERT by introducing sparse block structures into the attention matrix to reduce both memory consumption and training/inference time, which also enables attention heads to capture either short- or long-range contextual information.
Proceedings of the 1st Workshop on Representation Learning for NLP
TLDR
This paper applies layer-wise relevance propagation for the first time to natural language processing (NLP) and uses it to explain the predictions of a convolutional neural network trained on a topic categorization task.
UniK-QA: Unified Representations of Structured and Unstructured Knowledge for Open-Domain Question Answering
TLDR
It is demonstrated that the UniK-QA model is a simple and yet effective way to combine heterogeneous sources of knowledge, advancing the state-of-the-art results on two popular question answering benchmarks, NaturalQuestions and WebQuestions, by 3.5 and 2.6 points, respectively.
Unified Open-Domain Question Answering with Structured and Unstructured Knowledge
TLDR
This work homogenizes all sources by reducing them to text, and applies recent, powerful retriever-reader models which have so far been limited to text sources only to show that knowledge-base QA can be greatly improved when reformulated in this way.
Deep Learning in Natural Language Processing
TLDR
This chapter provides an introduction to the basics of natural language processing (NLP) as an integral part of artificial intelligence, and surveys the historical development of NLP, spanning over five decades, in terms of three waves.
Personalized Spam Filtering for Gray Mail
TLDR
This paper designs a light-weight user model that is highly scalable and can be easily combined with a traditional global spam filter and catches up to 40% more spam from gray mail in the low false-positive region.
Web-based Question Answering: Revisiting AskMSR
TLDR
This paper proposes a Web-QA system that removes the query pattern generation step and improves answer candidate generation and ranking steps, and outperforms AskMSR by 34 points of MRR.
Improve Spam Filtering by Detecting Gray Mail
TLDR
It is found that email campaigns that have messages labeled differently are the most reliable source for learning a gray mail detector, and a traditional statistical spam filter can still be improved consistently in different regions of the ROC curve by incorporating this new information.
...
...