NAACL2019论文

最近关注了一下NAACL2019，看了accepted papers，选了一些感兴趣的论文，有事没事看看，记录一下。

A Variational Approach to Weakly Supervised Document-Level Multi-Aspect Sentiment Classification     

Abusive Language Detection with Graph Convolutional Networks     

Adaptive Convolution for Text Classification     

Adversarial Category Alignment Network for Cross-domain Sentiment Classification     

An Effective Label Noise Model for DNN Text Classification     

An Embarrassingly Simple Approach for Transfer Learning from Pretrained Language Models 

Attention is not Explanation      

Convolutional Self-Attention Networks   

Dialogue Act Classification with Context-Aware Self-Attention       

Enhancing Opinion Role Labeling with Semantic-Aware Word Representations from Semantic Role Labeling     

Evaluating Text GANs as Language Models  

How Large A Vocabulary Does Text Classification Need? A Variational Approach on Vocabulary Selection ：

  文本分类任务中词表大小的选择，文章通过类似dropout的思想对每个词学习一个drop参数，通过阈值留下drop值小的词，选择最优的词表。

Incorporating Emoji Descriptions Improves Tweet Classification      

Integrating Semantic Knowledge to Tackle Zero-shot Text Classification     

Inter-Sentence Attention For Semantic Role Labeling     

Mitigating Uncertainty in Document Classification   

Partial Or Complete, That’s The Question    

Probabilistic Natural Language Generation with Wasserstein Autoencoders   

Ranking-Based AutoEncoder for Extreme Multi-label Classification   

Rethinking Complex Neural Network Architectures for Document Classification    

Syntax-aware Neural Semantic Role Labeling with Supertags          

Text Classification with Few Examples using Controlled Generalization         

A Radical‐aware Attention‐based Model for Chinese Text Classification    

What Is One Grain of Sand in the Desert? Analyzing Individual Neurons in Deep NLP Models   

Character‐Level Language Modeling with Deeper Self‐Attention   

Resisting Adversarial Attacks using Gaussian Mixture Variational Autoencoders  

Revisiting LSTM Networks for SemiS-upervised Text Classification via Mixed Objective Function 

InfoVAE: Balancing Learning and Inference in Variational Autoencoders   

Direct Training for Spiking Neural Networks: Faster, Larger, Better