Natural Language Model

Energy-Based Models for Text. arxiv

Dialog

Recipes for building an open-domain chatbot. arxiv

Generate Adversarial Model

FastWordBug: A Fast Method To Generate Adversarial Text Against NLP Applications. arxiv

Information Extraction

Rapid Adaptation of BERT for Information Extraction on Domain-Specific Business Documents. arxiv

Language Model

BERT-of-Theseus: Compressing BERT by Progressive Module Replacing. arxiv code
BERTweet: A pre-trained language model for English Tweets. arxiv code
Blank Language Models. arxiv
Controlling Computation versus Quality for Neural Sequence Models. arxiv
ELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators. arxiv code ⭐
Extending Multilingual BERT to Low-Resource Languages. arxiv
Limits of Detecting Text Generated by Large-Scale Language Models. arxiv
PALM: Pre-training an Autoencoding&Autoregressive Language Model for Context-conditioned Generation. arxiv
Pretrained Transformers Improve Out-of-Distribution Robustness. arxiv
Semantics-aware BERT for Language Understanding. arxiv code

Pos-tagging

Joint Embedding in Named Entity Linking on Sentence Level. arxiv

QA

AmbigQA: Answering Ambiguous Open-domain Questions. arxiv
Asking and Answering Questions to Evaluate the Factual Consistency of Summaries. arxiv
Conversational Question Answering over Passages by Leveraging Word Proximity Networks. arxiv code
Probing Emergent Semantics in Predictive Agents via Question Answering. arxiv
Unsupervised Commonsense Question Answering with Self-Talk. arxiv code

Text Classification

Light-Weighted CNN for Text Classification. arxiv
Multi-Label Text Classification using Attention-based Graph Neural Network. arxiv

Text Generation

NUBIA: NeUral Based Interchangeability Assessor for Text Generation. arxiv
Polarized-VAE: Proximity Based Disentangled Representation Learning for Text Generation. arxiv
Reverse Engineering Configurations of Neural Text Generation Models. arxiv