Enhanced LSTM for Natural Language Inference

Abstract

Reasoning and inference are central to human and artificial intelligence. Modeling inference in human language is very challenging. With the availability of large annotated data (Bowman et al., 2015), it has recently become feasible to train neural network based inference models, which have shown to be very effective. In this paper, we present a new state-of-the-art result, achieving the accuracy of 88.6% on the Stanford Natural Language Inference Dataset. Unlike the previous top models that use very complicated network architectures, we first demonstrate that carefully designing sequential inference models based on chain LSTMs can outperform all previous models. Based on this, we further show that by explicitly considering recursive architectures in both local inference modeling and inference composition, we achieve additional improvement. Particularly, incorporating syntactic parsing information contributes to our best result---it further improves the performance even when added to the already very strong model.

arXiv commentACL 2017

References in corpus (11)

• Adam: A Method for Stochastic Optimization
• Neural Machine Translation by Jointly Learning to Align and Translate
• Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
• Attention-Based Models for Speech Recognition
• Reasoning about Entailment with Neural Attention
• A large annotated corpus for learning natural language inference
• Learning Natural Language Inference using Bidirectional LSTM model and Inner-Attention
• Long Short-Term Memory-Networks for Machine Reading
• Order-Embeddings of Images and Language
• Exploring Question Understanding and Adaptation in Neural-Network-Based Question Answering
• A Neural Architecture Mimicking Humans End-to-End for Natural Language Inference

Cited by in corpus (242)

• Recent Trends in Deep Learning Based Natural Language Processing
• Generating Natural Adversarial Examples
• CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge
• Neural Natural Language Inference Models Enhanced with External Knowledge
• Adversarial Attacks on Deep Learning Models in Natural Language Processing: A Survey
• TabFact: A Large-scale Dataset for Table-based Fact Verification
• Natural Language Inference over Interaction Space
• BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions
• Evaluation of Text Generation: A Survey
• Visual Entailment: A Novel Task for Fine-Grained Image Understanding
• Bilateral Multi-Perspective Matching for Natural Language Sentences
• Neural Network Models for Paraphrase Identification, Semantic Textual Similarity, Natural Language Inference, and Question Answering
• Pretrained Transformers for Text Ranking: BERT and Beyond
• Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and Entailment
• A Large-Scale Corpus for Conversation Disentanglement
• Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inference
• R${}^3$: Reinforced Reader-Ranker for Open-Domain Question Answering
• Paradigm Shift in Natural Language Processing
• Neural Memory Networks for Seizure Type Classification
• Generating Persona Consistent Dialogues by Exploiting Natural Language Inference
• Distraction-Based Neural Networks for Document Summarization
• Recent Advances in Natural Language Inference: A Survey of Benchmarks, Resources, and Approaches
• Enhancing Sentence Embedding with Generalized Pooling
• Evaluation of Neural Architectures Trained with Square Loss vs Cross-Entropy in Classification Tasks
• Hybrid Tensor Decomposition in Neural Network Compression
• Attentive Deep Neural Networks for Legal Document Retrieval
• Encoded Summarization: Summarizing Documents into Continuous Vector Space for Legal Case Retrieval
• Using Prior Knowledge to Guide BERT's Attention in Semantic Textual Matching Tasks
• Adversarially Regularising Neural NLI Models to Integrate Logical Background Knowledge
• HellaSwag: Can a Machine Really Finish Your Sentence?
• SemMT: A Semantic-based Testing Approach for Machine Translation Systems
• Conflating point of interest (POI) data: A systematic review of matching methods
• Semantic Sentence Matching with Densely-connected Recurrent and Co-attentive Information
• Exploring Question Understanding and Adaptation in Neural-Network-Based Question Answering
• Shortcut-Stacked Sentence Encoders for Multi-Domain Inference
• CAIL2019-SCM: A Dataset of Similar Case Matching in Legal Domain
• Incorporating Domain Knowledge into Medical NLI using Knowledge Graphs
• Abductive Commonsense Reasoning
• Asking and Answering Questions to Evaluate the Factual Consistency of Summaries
• LOREN: Logic-Regularized Reasoning for Interpretable Fact Verification
• Transfer Fine-Tuning: A BERT Case Study
• Probing Biomedical Embeddings from Language Models
• Pair-Level Supervised Contrastive Learning for Natural Language Inference
• Sequential Attention-based Network for Noetic End-to-End Response Selection
• Don't Take the Easy Way Out: Ensemble Based Methods for Avoiding Known Dataset Biases
• Combining Fact Extraction and Verification with Neural Semantic Matching Networks
• GLoMo: Unsupervisedly Learned Relational Graphs as Transferable Representations
• Learning to Selectively Transfer: Reinforced Transfer Learning for Deep Text Matching
• Enhance word representation for out-of-vocabulary on Ubuntu dialogue corpus
• Towards a Robust Deep Neural Network in Texts: A Survey
• Explicit Contextual Semantics for Text Comprehension
• Dialog System Technology Challenge 7
• Multi-frame Feature Aggregation for Real-time Instrument Segmentation in Endoscopic Video
• On the Effectiveness of Low-Rank Matrix Factorization for LSTM Model Compression
• PubMedQA: A Dataset for Biomedical Research Question Answering
• BERT & Family Eat Word Salad: Experiments with Text Understanding
• Sentence Embeddings in NLI with Iterative Refinement Encoders
• Multi-Cast Attention Networks for Retrieval-based Question Answering and Response Prediction
• Neural Language Generation: Formulation, Methods, and Evaluation
• Fake News Detection as Natural Language Inference
• FarsTail: A Persian Natural Language Inference Dataset
• Adversarial attacks against Fact Extraction and VERification
• Natural Language Processing with Commonsense Knowledge: A Survey
• Solving Arithmetic Word Problems by Scoring Equations with Recursive Neural Networks
• TexSmart: A Text Understanding System for Fine-Grained NER and Enhanced Semantic Analysis
• Read + Verify: Machine Reading Comprehension with Unanswerable Questions
• On the Importance of Delexicalization for Fact Verification
• Improving Personality Consistency in Conversation by Persona Extending
• Multi-Level Matching and Aggregation Network for Few-Shot Relation Classification
• Unlearn Dataset Bias in Natural Language Inference by Fitting the Residual
• A comprehensive solution to retrieval-based chatbot construction
• Profile Consistency Identification for Open-domain Dialogue Agents
• DialBERT: A Hierarchical Pre-Trained Model for Conversation Disentanglement
• Show Your Work: Improved Reporting of Experimental Results
• Reasoning Over Semantic-Level Graph for Fact Checking
• Hierarchical Poset Decoding for Compositional Generalization in Language
• Lessons from Natural Language Inference in the Clinical Domain
• Getting To Know You: User Attribute Extraction from Dialogues
• Lightweight and Efficient Neural Natural Language Processing with Quaternion Networks
• To Tune or Not To Tune? How About the Best of Both Worlds?
• Improving Natural Language Inference with a Pretrained Parser
• Match-Prompt: Improving Multi-task Generalization Ability for Neural Text Matching via Prompt Learning
• Generate, Delete and Rewrite: A Three-Stage Framework for Improving Persona Consistency of Dialogue Generation
• PAWS: Paraphrase Adversaries from Word Scrambling
• Simple and Effective Text Matching with Richer Alignment Features
• Inter-BIN: Interaction-based Cross-architecture IoT Binary Similarity Comparison
• PAWS-X: A Cross-lingual Adversarial Dataset for Paraphrase Identification
• Infusing Finetuning with Semantic Dependencies
• Decision-making and control with diffractive optical networks
• Knowledge Enhanced Attention for Robust Natural Language Inference
• HIN: Hierarchical Inference Network for Document-Level Relation Extraction
• Generating Hierarchical Explanations on Text Classification via Feature Interaction Detection
• Universal Adversarial Attacks with Natural Triggers for Text Classification
• Legal Element-oriented Modeling with Multi-view Contrastive Learning for Legal Case Retrieval
• Interpreting Recurrent and Attention-Based Neural Models: a Case Study on Natural Language Inference
• Multi-turn Inference Matching Network for Natural Language Inference
• Interactive Matching Network for Multi-Turn Response Selection in Retrieval-Based Chatbots
• UKP-Athene: Multi-Sentence Textual Entailment for Claim Verification
• LGESQL: Line Graph Enhanced Text-to-SQL Model with Mixed Local and Non-Local Relations
• Tree Memory Networks for Modelling Long-term Temporal Dependencies
• "Going on a vacation" takes longer than "Going for a walk": A Study of Temporal Commonsense Understanding
• Augmenting semantic lexicons using word embeddings and transfer learning
• Breaking NLI Systems with Sentences that Require Simple Lexical Inferences
• Analyzing Compositionality-Sensitivity of NLI Models
• Learning Structured Text Representations
• SEED: Semantic Graph based Deep detection for type-4 clone
• Cross-Lingual Word Embedding Refinement by ${\ell }_1$ Norm Optimisation
• Filtered Inner Product Projection for Crosslingual Embedding Alignment
• Dynamic Multi-Level Multi-Task Learning for Sentence Simplification
• LadRa-Net: Locally-Aware Dynamic Re-read Attention Net for Sentence Semantic Matching
• Behavior Analysis of NLI Models: Uncovering the Influence of Three Factors on Robustness
• Reward Optimization for Neural Machine Translation with Learned Metrics
• Trimming and Improving Skip-thought Vectors
• A Sequential Matching Framework for Multi-turn Response Selection in Retrieval-based Chatbots
• Bilinear Fusion of Commonsense Knowledge with Attention-Based NLI Models
• Pipeline and Dataset Generation for Automated Fact-checking in Almost Any Language
• Dissecting Contextual Word Embeddings: Architecture and Representation
• GeoSQA: A Benchmark for Scenario-based Question Answering in the Geography Domain at High School Level
• Several Experiments on Investigating Pretraining and Knowledge-Enhanced Models for Natural Language Inference
• Building Sequential Inference Models for End-to-End Response Selection
• MAWSEO: Adversarial Wiki Search Poisoning for Illicit Online Promotion
• Effective and Imperceptible Adversarial Textual Attack via Multi-objectivization
• Weakly Supervised Explainable Phrasal Reasoning with Neural Fuzzy Logic
• AutoRC: Improving BERT Based Relation Classification Models via Architecture Search
• A Generate-Validate Approach to Answering Questions about Qualitative Relationships
• Parameter Re-Initialization through Cyclical Batch Size Schedules
• Answering while Summarizing: Multi-task Learning for Multi-hop QA with Evidence Extraction
• Investigating Biases in Textual Entailment Datasets
• Can neural networks understand monotonicity reasoning?
• Knowledge-incorporating ESIM models for Response Selection in Retrieval-based Dialog Systems
• Soft Label Memorization-Generalization for Natural Language Inference
• HELP: A Dataset for Identifying Shortcomings of Neural Models in Monotonicity Reasoning
• Cooperative Learning of Disjoint Syntax and Semantics
• A Zero Attentive Relevance Matching Networkfor Review Modeling in Recommendation System
• Improving Cross-Modal Alignment in Vision Language Navigation via Syntactic Information
• NeuralLog: Natural Language Inference with Joint Neural and Logical Reasoning
• Detecting and Explaining Causes From Text For a Time Series Event
• Perturbation-based Self-supervised Attention for Attention Bias in Text Classification
• How Large a Vocabulary Does Text Classification Need? A Variational Approach to Vocabulary Selection
• A Self-Training Method for Machine Reading Comprehension with Soft Evidence Extraction
• Neural CRF Model for Sentence Alignment in Text Simplification
• Learning Conceptual-Contextual Embeddings for Medical Text
• Improving Generalization by Incorporating Coverage in Natural Language Inference
• AlignVE: Visual Entailment Recognition Based on Alignment Relations
• Towards Robustifying NLI Models Against Lexical Dataset Biases
• Repurposing Entailment for Multi-Hop Question Answering Tasks
• Efficient Contextual Representation Learning Without Softmax Layer
• ICS-Assist: Intelligent Customer Inquiry Resolution Recommendation in Online Customer Service for Large E-Commerce Businesses
• An Empirical Study on Model-agnostic Debiasing Strategies for Robust Natural Language Inference
• Attentive Convolution: Equipping CNNs with RNN-style Attention Mechanisms
• Tag-Enhanced Tree-Structured Neural Networks for Implicit Discourse Relation Classification
• Automatic Validation of Textual Attribute Values in E-commerce Catalog by Learning with Limited Labeled Data
• Inducing and Embedding Senses with Scaled Gumbel Softmax
• Heterogeneous Network Embedding for Deep Semantic Relevance Match in E-commerce Search
• Machine Learning Meets Natural Language Processing -- The story so far
• What If We Simply Swap the Two Text Fragments? A Straightforward yet Effective Way to Test the Robustness of Methods to Confounding Signals in Nature Language Inference Tasks
• Effective Subword Segmentation for Text Comprehension
• Discrete Argument Representation Learning for Interactive Argument Pair Identification
• Mimic and Conquer: Heterogeneous Tree Structure Distillation for Syntactic NLP
• Modelling Domain Relationships for Transfer Learning on Retrieval-based Question Answering Systems in E-commerce
• Attention Boosted Sequential Inference Model
• Rewriter-Evaluator Architecture for Neural Machine Translation
• Interactive Model with Structural Loss for Language-based Abductive Reasoning
• Match${}^2$: A Matching over Matching Model for Similar Question Identification
• The Curse of Performance Instability in Analysis Datasets: Consequences, Source, and Suggestions
• Matching Text with Deep Mutual Information Estimation
• CoSQA: 20,000+ Web Queries for Code Search and Question Answering
• Topic-Aware Evidence Reasoning and Stance-Aware Aggregation for Fact Verification
• A Multi-Level Attention Model for Evidence-Based Fact Checking
• Promoting Diversity for End-to-End Conversation Response Generation
• DocNLI: A Large-scale Dataset for Document-level Natural Language Inference
• WikiCheck: An end-to-end open source Automatic Fact-Checking API based on Wikipedia
• Refining Raw Sentence Representations for Textual Entailment Recognition via Attention
• Dialogue Act Classification in Group Chats with DAG-LSTMs
• Imposing Label-Relational Inductive Bias for Extremely Fine-Grained Entity Typing
• Detecting Speaker Personas from Conversational Texts
• Fine Grained Human Evaluation for English-to-Chinese Machine Translation: A Case Study on Scientific Text
• Condition-Transforming Variational AutoEncoder for Conversation Response Generation
• Semantic Matching Against a Corpus: New Applications and Methods
• Alignment Attention by Matching Key and Query Distributions
• MuTual: A Dataset for Multi-Turn Dialogue Reasoning
• You Can Do Better! If You Elaborate the Reason When Making Prediction
• ForecastQA: A Question Answering Challenge for Event Forecasting with Temporal Text Data
• Attentive Tree-structured Network for Monotonicity Reasoning
• Empirical Evaluation of Multi-task Learning in Deep Neural Networks for Natural Language Processing
• Multi-turn Response Selection using Dialogue Dependency Relations
• A Hierarchical Deep Learning Natural Language Parser for Fashion
• Building an Efficient and Effective Retrieval-based Dialogue System via Mutual Learning
• KGSynNet: A Novel Entity Synonyms Discovery Framework with Knowledge Graph
• Pentagon at MEDIQA 2019: Multi-task Learning for Filtering and Re-ranking Answers using Language Inference and Question Entailment
• End-Task Oriented Textual Entailment via Deep Explorations of Inter-Sentence Interactions
• A multi-perspective combined recall and rank framework for Chinese procedure terminology normalization
• Evaluation Of Word Embeddings From Large-Scale French Web Content
• What's Missing: A Knowledge Gap Guided Approach for Multi-hop Question Answering
• Wasserstein Distance Regularized Sequence Representation for Text Matching in Asymmetrical Domains
• Answer Ranking for Product-Related Questions via Multiple Semantic Relations Modeling
• AWE: Asymmetric Word Embedding for Textual Entailment
• DEMN: Distilled-Exposition Enhanced Matching Network for Story Comprehension
• Program Enhanced Fact Verification with Verbalization and Graph Attention Network
• Discourse Marker Augmented Network with Reinforcement Learning for Natural Language Inference
• Inducing Alignment Structure with Gated Graph Attention Networks for Sentence Matching
• Learning to Decouple Relations: Few-Shot Relation Classification with Entity-Guided Attention and Confusion-Aware Training
• Utterance-to-Utterance Interactive Matching Network for Multi-Turn Response Selection in Retrieval-Based Chatbots
• Dropping Networks for Transfer Learning
• WikiContradiction: Detecting Self-Contradiction Articles on Wikipedia
• Review-guided Helpful Answer Identification in E-commerce
• Reliable Evaluations for Natural Language Inference based on a Unified Cross-dataset Benchmark
• L2R2: Leveraging Ranking for Abductive Reasoning
• Multi-Perspective Fusion Network for Commonsense Reading Comprehension
• Leveraging Linguistic Coordination in Reranking N-Best Candidates For End-to-End Response Selection Using BERT
• Cascaded Semantic and Positional Self-Attention Network for Document Classification
• Pre-trained Language Model Based Active Learning for Sentence Matching
• DGA-Net Dynamic Gaussian Attention Network for Sentence Semantic Matching
• Response Selection for Multi-Party Conversations with Dynamic Topic Tracking
• Answering Chinese Elementary School Social Study Multiple Choice Questions
• A Dialogue-based Information Extraction System for Medical Insurance Assessment
• Matching Questions and Answers in Dialogues from Online Forums
• Reasoning-Driven Question-Answering for Natural Language Understanding
• Improving Constituency Parsing with Span Attention
• Contextualized Embeddings based Convolutional Neural Networks for Duplicate Question Identification
• Unsupervised Pre-training with Structured Knowledge for Improving Natural Language Inference
• Multi-Perspective Inferrer: Reasoning Sentences Relationship from Holistic Perspective
• Interpretable Structure-aware Document Encoders with Hierarchical Attention
• On the Transferability of Adversarial Attacksagainst Neural Text Classifier
• Query-bag Matching with Mutual Coverage for Information-seeking Conversations in E-commerce
• Social Link Inference via Multi-View Matching Network from Spatio-Temporal Trajectories
• Don't be Contradicted with Anything! CI-ToD: Towards Benchmarking Consistency for Task-oriented Dialogue System
• MINIMAL: Mining Models for Data Free Universal Adversarial Triggers
• Generating Contradictory, Neutral, and Entailing Sentences
• Learning a Word-Level Language Model with Sentence-Level Noise Contrastive Estimation for Contextual Sentence Probability Estimation
• Teaching Syntax by Adversarial Distraction
• Personalizing Dialogue Agents via Meta-Learning
• Generalization in NLI: Ways (Not) To Go Beyond Simple Heuristics
• Does External Knowledge Help Explainable Natural Language Inference? Automatic Evaluation vs. Human Ratings
• BERT-DRE: BERT with Deep Recursive Encoder for Natural Language Sentence Matching
• Exploring End-to-End Differentiable Natural Logic Modeling
• R${}^2$-Net: Relation of Relation Learning Network for Sentence Semantic Matching
• Surf at MEDIQA 2019: Improving Performance of Natural Language Inference in the Clinical Domain by Adopting Pre-trained Language Model
• Discriminatively-Tuned Generative Classifiers for Robust Natural Language Inference
• Visuallly Grounded Generation of Entailments from Premises
• Find a Reasonable Ending for Stories: Does Logic Relation Help the Story Cloze Test?
• Few Shot Learning for Information Verification