总结 | ACL2022主会论文分类整理（二）

表示学习

• A Contrastive Framework for Learning Sentence Representations from Pairwise and Triple-wise Perspective in Angular Space

• Auto-Debias: Debiasing Masked Language Models with Automated Biased Prompts

• Compact Token Representations with Contextual Quantization for Efficient Document Re-ranking

• Contextual Representation Learning beyond Masked Language Modeling

• Contrastive Visual Semantic Pretraining Magnifies the Semantics of Natural Language Representations

• Cross-Lingual Contrastive Learning for Fine-Grained Entity Typing for LowResource Languages

• Cross-Modal Discrete Representation Learning

• Debiased Contrastive Learning of Unsupervised Sentence Representations

• Enhancing Chinese Pre-trained Language Model via Heterogeneous Linguistics Graph

• GL-CLeF: A Global–Local Contrastive Learning Framework for Crosslingual Spoken Language Understanding

• Improving Event Representation via Simultaneous Weakly Supervised Contrastive Learning and Clustering

• Just Rank: Rethinking Evaluation with Word and Sentence Similarities

• Language-agnostic BERT Sentence Embedding

• Learning Disentangled Representations of Negation and Uncertainty

• Learning Disentangled Textual Representations via Statistical Measures of Similarity

• Multilingual Molecular Representation Learning via Contrastive Pre-training

• Nibbling at the Hard Core of Word Sense Disambiguation

• Noisy Channel Language Model Prompting for Few-Shot Text Classification

• Rare and Zero-shot Word Sense Disambiguation using Z-Reweighting

• Sentence-level Privacy for Document Embeddings

• Softmax Bottleneck Makes Language Models Unable to Represent Multimode Word Distributions

• SPoT: Better Frozen Model Adaptation through Soft Prompt Transfer

• Tackling Fake News Detection by Continually Improving Social Context Representations using Graph Neural Networks

• The Grammar-Learning Trajectories of Neural Language Models

• Using Context-to-Vector with Graph Retrofitting to Improve Word Embeddings

Machine Learning for NLP【NLP中的机器学习】

• A Rationale-Centric Framework for Human-in-the-loop Machine Learning

• Bias Mitigation in Machine Translation Quality Estimation

• Disentangled Sequence to Sequence Learning for Compositional Generalization

• DoCoGen: Domain Counterfactual Generation for Low Resource Domain Adaptation

• Domain Adaptation in Multilingual and Multi-Domain Monolingual Settings for Complex Word Identification

• Domain Knowledge Transferring for Pre-trained Language Model via Calibrated Activation Boundary Distillation

• Learning Functional Distributional Semantics with Visual Data

• Leveaging Relaxed Equilibrium by Lazy Transition for Sequence Modeling

• Local Languages, Third Spaces, and other High-Resource Scenarios

• Meta-learning via Language Model In-context Tuning

• MPII: Multi-Level Mutual Promotion for Inference and Interpretation

• On the Calibration of Pre-trained Language Models using Mixup Guided by Area Under the Margin and Saliency

• Overcoming a Theoretical Limitation of Self-Attention

• Rethinking Negative Sampling for Handling Missing Entity Annotations

• Rethinking Self-Supervision Objectives for Generalizable Coherence Modeling

• Robust Lottery Tickets for Pre-trained Language Models

• Sharpness-Aware Minimization Improves Language Model Generalization

• Skill Induction and Planning with Latent Language

• The Trade-offs of Domain Adaptation for Neural Language Models

• Distributionally Robust Finetuning BERT for Covariate Drift in Spoken Language Understanding

• Learning to Imagine: Integrating Counterfactual Thinking in Neural Discrete Reasoning

Machine Translation and Multilinguality【机器翻译与多语】

翻译

• Alignment-Augmented Consistent Translation for Multilingual Open Information Extraction

• Alternative Input Signals Ease Transfer in Multilingual Machine Translation

• BiTIIMT: A Bilingual Text-infilling Method for Interactive Machine Translation

• Bridging the Data Gap between Training and Inference for Unsupervised Neural Machine Translation

• Can Transformer be Too Compositional? Analysing Idiom Processing in Neural Machine Translation

• CipherDAug: Ciphertext based Data Augmentation for Neural Machine Translation

• Conditional Bilingual Mutual Information Based Adaptive Training for Neural Machine Translation

• Confidence Based Bidirectional Global Context Aware Training Framework for Neural Machine Translation

• DEEP: DEnoising Entity Pre-training for Neural Machine Translation

• DiBiMT: A Novel Benchmark for Measuring Word Sense Disambiguation Biases in Machine Translation

• Divide and Rule: Effective Pre-Training for Context-Aware Multi-Encoder Translation Models

• EAG: Extract and Generate Multi-way Aligned Corpus for Complete Multilingual Neural Machine Translation

• Efficient Cluster-Based k-Nearest-Neighbor Machine Translation

• Flow-Adapter Architecture for Unsupervised Machine Translation

• From Simultaneous to Streaming Machine Translation by Leveraging Streaming History

• Improving Word Translation via Two-Stage Contrastive Learning

• Integrating Vectorized Lexical Constraints for Neural Machine Translation

• Investigating Failures of Automatic Translation in the Case of Unambiguous Gender

• Learning Adaptive Segmentation Policy for End-to-End Simultaneous Translation

• Learning Confidence for Transformer-based Neural Machine Translation

• Learning to Generalize to More: Continuous Semantic Augmentation for Neural Machine Translation

• Learning When to Translate for Streaming Speech

• Measuring and Mitigating Name Biases in Neural Machine Translation

• Modeling Dual Read/Write Paths for Simultaneous Machine Translation

• MSP: Multi-Stage Prompting for Making Pre-trained Language Models Better Translators

• Multilingual Document-Level Translation Enables Zero-Shot Transfer From Sentences to Documents

• Multilingual Mix: Example Interpolation Improves Multilingual Neural Machine Translation

• Neural Machine Translation with Phrase-Level Universal Visual Representations

• On Vision Features in Multimodal Machine Translation

• Overcoming Catastrophic Forgetting beyond Continual Learning: Balanced Training for Neural Machine Translation

• Prediction Difference Regularization against Perturbation for Neural Machine Translation

• Redistributing Low-Frequency Words: Making the Most of Monolingual Data in Non-Autoregressive Translation

• Reducing Position Bias in Simultaneous Machine Translation with Length Aware Framework

• Scheduled Multi-task Learning for Neural Chat Translation

• The Paradox of the Compositionality of Natural Language: A Neural Machine Translation Case Study

• Towards Making the Most of Cross-Lingual Transfer for Zero-Shot Neural Machine Translation

• Understanding and Improving Sequence-to-Sequence Pretraining for Neural Machine Translation

• Unified Speech-Text Pre-training for Speech Translation and Recognition

• UniTE: Unified Translation Evaluation

• Universal Conditional Masked Language Pre-training for Neural Machine Translation

多语

• AmericasNLI: Evaluating Zero-shot Natural Language Understanding of Pretrained Multilingual Models in Truly Low-resource Languages

• Cross-Lingual Ability of Multilingual Masked Language Models: A Study of Language Structure

• Domain Adaptation in Multilingual and Multi-Domain Monolingual Settings for Complex Word Identification

• Expanding Pretrained Models to Thousands More Languages via Lexiconbased Adaptation

• Match the Script, Adapt if Multilingual: Analyzing the Effect of Multilingual Pretraining on Cross-lingual Transferability

• mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models

• Multi Task Learning For Zero Shot Performance Prediction of Multilingual Models

• Multilingual Generative Language Models for Zero-Shot Cross-Lingual Event Argument Extraction

• Multilingual Knowledge Graph Completion with Self-Supervised Adaptive Graph Alignment

• Multilingual Molecular Representation Learning via Contrastive Pre-training

• Multilingual unsupervised sequence segmentation transfers to extremely low-resource languages

• One Country, 700+ Languages: NLP Challenges for Underrepresented Languages and Dialects in Indonesia

• Prix-LM: Pretraining for Multilingual Knowledge Base Construction

• Probing Structured Pruning on Multilingual Pre-trained Models: Settings, Algorithms, and Efficiency

Question Answering【问答与理解】

阅读理解

• AdaLoGN: Adaptive Logic Graph Network for Reasoning-Based Machine Reading Comprehension

• Deep Inductive Logic Reasoning for Multi-Hop Reading Comprehension

• Improving Machine Reading Comprehension with Contextualized Commonsense Knowledge

• Learning Disentangled Semantic Representations for Zero-Shot CrossLingual Transfer in Multilingual Machine Reading Comprehension

• Lite Unified Modeling for Discriminative Reading Comprehension

• Modeling Temporal-Modal Entity Graph for Procedural Multimodal Machine Comprehension

• What Makes Reading Comprehension Questions Difficult?

• MultiHiertt: Numerical Reasoning over Multi Hierarchical Tabular and Textual Data

问答

• Answer-level Calibration for Free-form Multiple Choice Question Answering

• Answering Open-Domain Multi-Answer Questions via a Recall-then-Verify Framework

• CQG: A Simple and Effective Controlled Generation Framework for Multihop Question Generation

• Ditch the Gold Standard: Re-evaluating Conversational Question Answering

• Generated Knowledge Prompting for Commonsense Reasoning

• How Do We Answer Complex Questions: Discourse Structure of Long-form Answers

• Hypergraph Transformer: Weakly-Supervised Multi-hop Reasoning for Knowledge-based Visual Question Answering

• Hyperlink-induced Pre-training for Passage Retrieval in Open-domain Question Answering

• Improving Time Sensitivity for Question Answering over Temporal Knowledge Graphs

• It is AI’s Turn to Ask Humans a Question: Question-Answer Pair Generation for Children’s Story Books

• KaFSP: Knowledge-Aware Fuzzy Semantic Parsing for Conversational Question Answering over a Large-Scale Knowledge Base

• KG-FiD: Infusing Knowledge Graph in Fusion-in-Decoder for Open-Domain Question Answering

• MMCoQA: Conversational Question Answering over Text, Tables, and Images

• Modeling Multi-hop Question Answering as Single Sequence Prediction

• On the Robustness of Question Rewriting Systems to Questions of Varying Hardness

• Open Domain Question Answering with A Unified Knowledge Interface

• Program Transfer for Answering Complex Questions over Knowledge Bases

• Retrieval-guided Counterfactual Generation for QA

• RNG-KBQA: Generation Augmented Iterative Ranking for Knowledge Base Question Answering

• Sequence-to-Sequence Knowledge Graph Completion and Question Answering

• Simulating Bandit Learning from User Feedback for Extractive Question Answering

• Subgraph Retrieval Enhanced Model for Multi-hop Knowledge Base Question Answering

• Synthetic Question Value Estimation for Domain Adaptation of Question Answering

• Your Answer is Incorrect… Would you like to know why? Introducing a Bilingual Short Answer Feedback Dataset

Resources and Evaluation【数据集与评估方法】

数据集

• A Statutory Article Retrieval Dataset in French

• CBLUE: A Chinese Biomedical Language Understanding Evaluation Benchmark

• Chart-to-Text: A Large-Scale Benchmark for Chart Summarization

• CICERO: A Dataset for Contextualized Commonsense Inference in Dialogues

• CLUES: A Benchmark for Learning Classifiers using Natural Language Explanations

• ConditionalQA: A Complex Reading Comprehension Dataset with Conditional Answers

• Cree Corpus: A Collection of nêhiyawêwin Resources

• Detecting Unassimilated Borrowings in Spanish: An Annotated Corpus and Approaches to Modeling

• DialFact: A Benchmark for Fact-Checking in Dialogue

• DiBiMT: A Novel Benchmark for Measuring Word Sense Disambiguation Biases in Machine Translation

• Down and Across: Introducing Crossword-Solving as a New NLP Benchmark

• e-CARE: a New Dataset for Exploring Explainable Causal Reasoning

• EntSUM: A Data Set for Entity-Centric Extractive Summarization

• ePiC: Employing Proverbs in Context as a Benchmark for Abstract Language Understanding

• FairLex: A Multilingual Benchmark for Evaluating Fairness in Legal Text Processing

• Fantastic Questions and Where to Find Them: FairytaleQA – An Authentic Dataset for Narrative Comprehension

• Few-Shot Tabular Data Enrichment Using Fine-Tuned Transformer Architectures

• French CrowS-Pairs: Extending a challenge dataset for measuring social bias in masked language models to a language other than English

• From text to talk: Harnessing conversational corpora for humane and diversity-aware language technology

• HiTab: A Hierarchical Table Dataset for Question Answering and Natural Language Generation

• IAM: A Comprehensive and Large-Scale Dataset for Integrated Argument Mining Tasks

• Image Retrieval from Contextual Descriptions

• KQA Pro: A Dataset with Explicit Compositional Programs for Complex Question Answering over Knowledge Base

• LexGLUE: A Benchmark Dataset for Legal Language Understanding in English

• M3ED: Multi-modal Multi-scene Multi-label Emotional Dialogue Database

• MSCTD: A Multimodal Sentiment Chat Translation Dataset

• NumGLUE: A Suite of Fundamental yet Challenging Mathematical Reasoning Tasks

• QuoteR: A Benchmark of Quote Recommendation for Writing

• Reports of personal experiences and stories in argumentation: datasets and analysis

• RNSum: A Large-Scale Dataset for Automatic Release Note Generation via Commit Logs Summarization

• SciNLI: A Corpus for Natural Language Inference on Scientific Text

• SummScreen: A Dataset for Abstractive Screenplay Summarization

• SUPERB-SG: Enhanced Speech processing Universal PERformance Benchmark for Semantic and Generative Capabilities

• Textomics: A Dataset for Genomics Data Summary Generation

• The Moral Integrity Corpus: A Benchmark for Ethical Dialogue Systems

• ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection

• VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena

• WatClaimCheck: A new Dataset for Claim Entailment and Inference

• Your Answer is Incorrect… Would you like to know why? Introducing a Bilingual Short Answer Feedback Dataset