EMNLP（Conference on Empirical Methods in Natural Language Processing）是自然语言处理领域的顶会，由ACL学会下的语言数据特殊兴趣小组SIGDAT组织，自1996年起每年举行。

自然语言处理顶会 EMNLP 2018 多样化地在更有难度的任务的尝试、对已有方法的更深入的探讨、对新方向的探索中展现了计算语言学的更多可能。Facebook 和谷歌一如既往地在 EMNLP 2018 上收获颇丰。Facebook 有 14 篇论文被 EMNLP 接受，谷歌则有多达 26 篇（有一篇是两者合作完成）。虽然目前 EMNLP 还没有正式宣布论文接受数目，但我们预计这个数目大概会在四百多篇。这样一来，有接近 10% 的收录论文就直接被 Facebook 和谷歌「承包」了。而且，EMNLP 2018 的两篇最佳长论文也刚好一篇出自 Facebook，一篇出自谷歌。可见 https://arxiv.org/abs/1804.07755 与 http://aclweb.org/anthology/D18-1548。

【Facebook论文清单】

D18-1117
A Dataset for Telling the Stories of Social Media Videos
一个从社交媒体视频讲故事的数据集
http://aclweb.org/anthology/D18-1117

D18-1181
Auto-Encoding Dictionary Definitions into Consistent Word Embeddings
把自动编码字典的定义转换为连续的词嵌入
http://aclweb.org/anthology/D18-1181

D18-1128
Do explanations make VQA models more predictable to a human?
视觉问答模型的解释能让它们对人类来说更好预测吗？
http://aclweb.org/anthology/D18-1128

D18-1176
Dynamic Meta-Embeddings for Improved Sentence Representations
用动态元嵌入提供更好的句子表征
http://aclweb.org/anthology/D18-1176

D18-1073
Extending Neural Generative Conversational Model using External Knowledge Sources
用外部知识源拓展神经生成式对话模型
http://aclweb.org/anthology/D18-1073

D18-1119
How agents see things: On visual representations in an emergent language game
智能体是怎么观察事物的：对演变语言游戏中的视觉表征的研究
http://aclweb.org/anthology/D18-1119

D18-1330
Loss in Translation: Learning Bilingual Word Mapping with a Retrieval Criterion
在翻译中迷失：通过回溯条件学习双语单词映射
http://aclweb.org/anthology/D18-1330

D18-1239
Neural Compositional Denotational Semantics for Question Answering
用于问答的神经合成标志语义
http://aclweb.org/anthology/D18-1239

D18-1043
Non-Adversarial Unsupervised Word Translation
非对抗性训练的无监督单词翻译
http://aclweb.org/anthology/D18-1043

D18-1549
Phrase-Based & Neural Unsupervised Machine Translation
基于词语的、无监督神经网络的机器翻译
http://aclweb.org/anthology/D18-1549

D18-1300
Semantic Parsing for Task Oriented Dialog using Hierarchical Representations
通过层次化表征对面向任务的对话做语义解析
http://aclweb.org/anthology/D18-1300

D18-1298
Training Millions of Personalized Dialogue Agents
训练一百万个个性化定制的对话机器人
http://aclweb.org/anthology/D18-1298

D18-1045
Understanding Back-Translation at Scale
理解大规模反向翻译
与谷歌合作完成
http://aclweb.org/anthology/D18-1045

D18-1269
XNLI: Evaluating Cross-lingual Sentence Representations
XNLI：评价跨语言的句子表征
http://aclweb.org/anthology/D18-1269

【Google论文清单】

D18-1277
​A Challenge Set and Methods for Noun-Verb Ambiguity
非动词单词歧义性的挑战性数据集及方法
http://aclweb.org/anthology/D18-1277

D18-1030
A Fast, Compact, Accurate Model for Language Identification of Codemixed Text
一个快速、紧凑、准确的模型，用于混有代码的文本的语言辨别
http://aclweb.org/anthology/D18-1030

D18-1419
AirDialogue: An Environment for Goal-Oriented Dialogue Research
AirDialogue：一个为面向目标的对话研究设计的环境
http://aclweb.org/anthology/D18-1419

D18-1374
Content Explorer: Recommending Novel Entities for a Document Writer
内容浏览器：为文档作者推荐新的实体
http://aclweb.org/anthology/D18-1374

D18-1211
Deep Relevance Ranking using Enhanced Document-Query Interactions
借助增强的文档-查询互动的深度关联性排序
http://aclweb.org/anthology/D18-1211

D18-1259
HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering
HotpotQA：用于多样化、可解释的多次跳转问答的数据集
http://aclweb.org/anthology/D18-1259

D18-1091
Identifying Well-formed Natural Language Questions
识别格式完备的自然语言问题
http://aclweb.org/anthology/D18-1091

D18-1080
Learning To Split and Rephrase From Wikipedia Edit History
从维基百科的编辑历史中学习分割和重新组织语句
http://aclweb.org/anthology/D18-1080

D18-1548
Linguistically-Informed Self-Attention for Semantic Role Labeling
用于语义角色标注的考虑语言学信息的自我注意力方法
http://aclweb.org/anthology/D18-1548

D18-1455
Open Domain Question Answering Using Early Fusion of Knowledge Bases and Text
使用知识库和文本早期聚合技术的开放领域问答
http://aclweb.org/anthology/D18-1455

D18-1405
Noise Contrastive Estimation for Conditional Models: Consistency and Statistical Efficiency
条件模型的噪声对比估计：连续性与统计效率研究
http://aclweb.org/anthology/D18-1405

D18-1347
Part-of-Speech Tagging for Code-Switched, Transliterated Texts without Explicit Language Identification
在不显式指定语言的状况下对代码转换的、根据读音拼写的文本做部分语句标注
http://aclweb.org/anthology/D18-1347

D18-1052
Phrase-Indexed Question Answering: A New Challenge for Scalable Document Comprehension
以短语为索引的问答：可拓展文档理解的新挑战
http://aclweb.org/anthology/D18-1052

D18-1266
Policy Shaping and Generalized Update Equations for Semantic Parsing from Denotations
用于标志文本的语意解析的策略塑造以及通用化更新公式
http://aclweb.org/anthology/D18-1266

D18-1461
Revisiting Character-Based Neural Machine Translation with Capacity and Compression
带着容量和压缩的疑问，重新思考基于字符的神经机器翻译
http://aclweb.org/anthology/D18-1461

D18-1105
Self-governing neural networks for on-device short text classification
用于终端设备上的短文本分类任务的自组织神经网络
http://aclweb.org/anthology/D18-1105

D18-1217
Semi-Supervised Sequence Modeling with Cross-View Training
带有跨视角训练的半监督序列建模
http://aclweb.org/anthology/D18-1217

D18-1529
State-of-the-art Chinese Word Segmentation with Bi-LSTMs
利用双向LSTM的顶级中文分词
http://aclweb.org/anthology/D18-1529

D18-1253
Subgoal Discovery for Hierarchical Dialogue Policy Learning
用于层次化对话策略学习的次目标发掘
http://aclweb.org/anthology/D18-1253

D18-1100
SwitchOut: an Efficient Data Augmentation Algorithm for Neural Machine Translation
SwitchOut：一个用于神经机器翻译的高效的数据扩增算法
http://aclweb.org/anthology/D18-1100

D18-1324
The Importance of Generation Order in Language Modeling
对语言建模中生成顺序的重要性的探究
http://aclweb.org/anthology/D18-1324

D18-1338
Training Deeper Neural Machine Translation Models with Transparent Attention
利用透明注意力训练更深的神经机器翻译模型
http://aclweb.org/anthology/D18-1338

D18-1045
Understanding Back-Translation at Scale
理解大规模反向翻译
http://aclweb.org/anthology/D18-1045

D18-1426
Unsupervised Natural Language Generation with Denoising Autoencoders
利用降噪自动编码器进行无监督自然语言生成
http://aclweb.org/anthology/D18-1426

D18-1028
WikiAtomicEdits: A Multilingual Corpus of Wikipedia Edits for Modeling Language and Discourse
WikiAtomicEdits：用于语言建模和论述的多语言维基百科编辑记录语料库
http://aclweb.org/anthology/D18-1028

D18-1305
WikiConv: A Corpus of the Complete Conversational History of a Large Online Collaborative Community
WikiConv：一个大型在线合作型社区的完整对话历史语料库
http://aclweb.org/anthology/D18-1305