Convolutional Neural Networks for Sentence Classification

Abstract

We report on a series of experiments with convolutional neural networks (CNN) trained on top of pre-trained word vectors for sentence-level classification tasks. We show that a simple CNN with little hyperparameter tuning and static vectors achieves excellent results on multiple benchmarks. Learning task-specific vectors through fine-tuning offers further gains in performance. We additionally propose a simple modification to the architecture to allow for the use of both task-specific and static vectors. The CNN models discussed herein improve upon the state of the art on 4 out of 7 tasks, which include sentiment analysis and question classification.

To appear in EMNLP 2014

References in corpus (4)

• Improving neural networks by preventing co-adaptation of feature detectors
• ADADELTA: An Adaptive Learning Rate Method
• Seeing stars: Exploiting class relationships for sentiment categorization with respect to rating scales
• A Convolutional Neural Network for Modelling Sentences

Cited by in corpus (141)

• Deep Learning for Anomaly Detection: A Survey
• FPGA-based Accelerators of Deep Learning Networks for Learning and Classification: A Review
• Adversarial Attacks on Deep Learning Models in Natural Language Processing: A Survey
• Hierarchical Multiscale Recurrent Neural Networks
• DKN: Deep Knowledge-Aware Network for News Recommendation
• Deep Learning and Its Applications to Machine Health Monitoring: A Survey
• dna2vec: Consistent vector representations of variable-length k-mers
• Emotion Detection on TV Show Transcripts with Sequence-based Convolutional Neural Networks
• ATRank: An Attention-Based User Behavior Modeling Framework for Recommendation
• Joint Deep Modeling of Users and Items Using Reviews for Recommendation
• Automatic Rule Extraction from Long Short Term Memory Networks
• Discriminative Neural Sentence Modeling by Tree-Based Convolution
• Improving the Accuracy of Pre-trained Word Embeddings for Sentiment Analysis
• Interpretation of Prediction Models Using the Input Gradient
• A Survey on Transfer Learning in Natural Language Processing
• Learning to Decode for Future Success
• Neural Information Retrieval: A Literature Review
• Neural Models for Sequence Chunking
• Exploring Question Understanding and Adaptation in Neural-Network-Based Question Answering
• Deep-Sentiment: Sentiment Analysis Using Ensemble of CNN and Bi-LSTM Models
• Deep Multi-Task Learning with Shared Memory
• Visualizing and Understanding Curriculum Learning for Long Short-Term Memory Networks
• Towards Scalable and Reliable Capsule Networks for Challenging NLP Applications
• Word Embeddings and Their Use In Sentence Classification Tasks
• Knowledge as a Teacher: Knowledge-Guided Structural Attention Networks
• Machine Reading Comprehension: a Literature Review
• Deepr: A Convolutional Net for Medical Records
• A Survey on Knowledge Graph-Based Recommender Systems
• A Neural Topic-Attention Model for Medical Term Abbreviation Disambiguation
• Learning Loss Functions for Semi-supervised Learning via Discriminative Adversarial Networks
• DOC: Deep Open Classification of Text Documents
• Open-Domain Targeted Sentiment Analysis via Span-Based Extraction and Classification
• CrisisBERT: a Robust Transformer for Crisis Classification and Contextual Crisis Embedding
• Misogynistic Tweet Detection: Modelling CNN with Small Datasets
• Recurrent Neural Network-Based Sentence Encoder with Gated Attention for Natural Language Inference
• Joint Online Spoken Language Understanding and Language Modeling with Recurrent Neural Networks
• Highrisk Prediction from Electronic Medical Records via Deep Attention Networks
• Few-shot Learning with Meta Metric Learners
• Combining Convolution and Recursive Neural Networks for Sentiment Analysis
• Searching for Effective Neural Extractive Summarization: What Works and What's Next
• KATE: K-Competitive Autoencoder for Text
• Comparing Rule-Based and Deep Learning Models for Patient Phenotyping
• Multimodal Attribute Extraction
• Multichannel CNN with Attention for Text Classification
• Deep Learning and Word Embeddings for Tweet Classification for Crisis Response
• Character-level Convolutional Network for Text Classification Applied to Chinese Corpus
• Gated Recurrent Neural Network Approach for Multilabel Emotion Detection in Microblogs
• Optimizing Gross Merchandise Volume via DNN-MAB Dynamic Ranking Paradigm
• Integrating Semantic Knowledge to Tackle Zero-shot Text Classification
• From Independent Prediction to Re-ordered Prediction: Integrating Relative Position and Global Label Information to Emotion Cause Identification
• Exploiting Sentence and Context Representations in Deep Neural Models for Spoken Language Understanding
• Generating lyrics with variational autoencoder and multi-modal artist embeddings
• Conditional BERT Contextual Augmentation
• A Data-driven Neural Network Architecture for Sentiment Analysis
• Syntactic Recurrent Neural Network for Authorship Attribution
• Lyrics-Based Music Genre Classification Using a Hierarchical Attention Network
• Generative Encoder-Decoder Models for Task-Oriented Spoken Dialog Systems with Chatting Capability
• $k$-Nearest Neighbor Augmented Neural Networks for Text Classification
• Neural Contextual Conversation Learning with Labeled Question-Answering Pairs
• Modeling Long-Range Context for Concurrent Dialogue Acts Recognition
• OCC: A Smart Reply System for Efficient In-App Communications
• Knowledge-Enriched Transformer for Emotion Detection in Textual Conversations
• OCoR: An Overlapping-Aware Code Retriever
• Dynamic Compositional Neural Networks over Tree Structure
• Predicting US State-Level Agricultural Sentiment as a Measure of Food Security with Tweets from Farming Communities
• Weak Supervision for Fake News Detection via Reinforcement Learning
• Image Transformation can make Neural Networks more robust against Adversarial Examples
• Label-guided Learning for Text Classification
• Leveraging Multi-Source Weak Social Supervision for Early Detection of Fake News
• Generation-Distillation for Efficient Natural Language Understanding in Low-Data Settings
• CAT-Gen: Improving Robustness in NLP Models via Controlled Adversarial Text Generation
• Controlled CNN-based Sequence Labeling for Aspect Extraction
• Visual and Textual Sentiment Analysis Using Deep Fusion Convolutional Neural Networks
• A Sequential Matching Framework for Multi-turn Response Selection in Retrieval-based Chatbots
• Learning Phrase Embeddings from Paraphrases with GRUs
• Computation Error Analysis of Block Floating Point Arithmetic Oriented Convolution Neural Network Accelerator Design
• Contextual Text Style Transfer
• Online Deception Detection Refueled by Real World Data Collection
• Light-Weighted CNN for Text Classification
• Phonetic-enriched Text Representation for Chinese Sentiment Analysis with Reinforcement Learning
• What Do You Mean I'm Funny? Personalizing the Joke Skill of a Voice-Controlled Virtual Assistant
• CatGAN: Category-aware Generative Adversarial Networks with Hierarchical Evolutionary Learning for Category Text Generation
• Exploiting Synchronized Lyrics And Vocal Features For Music Emotion Detection
• Real-Time Steganalysis for Stream Media Based on Multi-channel Convolutional Sliding Windows
• Data Sanity Check for Deep Learning Systems via Learnt Assertions
• CMU LiveMedQA at TREC 2017 LiveQA: A Consumer Health Question Answering System
• Learning Semantically Coherent and Reusable Kernels in Convolution Neural Nets for Sentence Classification
• MALCOM: Generating Malicious Comments to Attack Neural Fake News Detection Models
• CANDiS: Coupled & Attention-Driven Neural Distant Supervision
• Predicting Helpfulness of Online Reviews
• ST${}^2$: Small-data Text Style Transfer via Multi-task Meta-Learning
• Learning Effective Representations for Person-Job Fit by Feature Fusion
• Contextual and Position-Aware Factorization Machines for Sentiment Classification
• Sparse Lifting of Dense Vectors: Unifying Word and Sentence Representations
• Predicting Discharge Medications at Admission Time Based on Deep Learning
• Learning Invariant Representations of Social Media Users
• Cell Type Identification from Single-Cell Transcriptomic Data via Semi-supervised Learning
• An Investigation of Transfer Learning-Based Sentiment Analysis in Japanese
• Water Preservation in Soan River Basin using Deep Learning Techniques
• Predicting user intent from search queries using both CNNs and RNNs
• MoodSwipe: A Soft Keyboard that Suggests Messages Based on User-Specified Emotions
• Combine Convolution with Recurrent Networks for Text Classification
• Combination of multiple Deep Learning architectures for Offensive Language Detection in Tweets
• Sentiment Classification with Word Attention based on Weakly Supervised Learning with a Convolutional Neural Network
• Predicting Missing Information of Key Aspects in Vulnerability Reports
• Product Function Need Recognition via Semi-supervised Attention Network
• Impact of Feature Selection on Micro-Text Classification
• DeepChrome: Deep-learning for predicting gene expression from histone modifications
• SLAM-Inspired Simultaneous Contextualization and Interpreting for Incremental Conversation Sentences
• Ranking-Based Autoencoder for Extreme Multi-label Classification
• Medical Multimodal Classifiers Under Scarce Data Condition
• Hybrid Tiled Convolutional Neural Networks for Text Sentiment Classification
• Embedding Compression with Isotropic Iterative Quantization
• Semi-supervised Text Style Transfer: Cross Projection in Latent Space
• PERS: A Personalized and Explainable POI Recommender System
• Text Classification based on Multi-granularity Attention Hybrid Neural Network
• Graph Convolution for Multimodal Information Extraction from Visually Rich Documents
• A Multichannel Convolutional Neural Network For Cross-language Dialog State Tracking
• Entropy-Based Modeling for Estimating Soft Errors Impact on Binarized Neural Network Inference
• Unsupervised and Supervised Structure Learning for Protein Contact Prediction
• How to Evaluate Word Representations of Informal Domain?
• Large-scale Gender/Age Prediction of Tumblr Users
• SmartBullets: A Cloud-Assisted Bullet Screen Filter based on Deep Learning
• Do users talk about the software in my product? Analyzing user reviews on IoT products
• Using Context Events in Neural Network Models for Event Temporal Status Identification
• Deep Learning based, end-to-end metaphor detection in Greek language with Recurrent and Convolutional Neural Networks
• DrugDBEmbed : Semantic Queries on Relational Database using Supervised Column Encodings
• EventMapper: Detecting Real-World Physical Events Using Corroborative and Probabilistic Sources
• End-to-end Learning for Short Text Expansion
• CRNN: A Joint Neural Network for Redundancy Detection
• Conversational Help for Task Completion and Feature Discovery in Personal Assistants
• Deep Unified Multimodal Embeddings for Understanding both Content and Users in Social Media Networks
• Search Intelligence: Deep Learning For Dominant Category Prediction
• Batch-Augmented Multi-Agent Reinforcement Learning for Efficient Traffic Signal Optimization
• Towards end-to-end pulsed eddy current classification and regression with CNN
• A machine learning method correlating pulse pressure wave data with pregnancy
• Towards Controlled Transformation of Sentiment in Sentences
• A Deep Learning-Based Approach for Measuring the Domain Similarity of Persian Texts
• OhioState at IJCNLP-2017 Task 4: Exploring Neural Architectures for Multilingual Customer Feedback Analysis
• Stylized Text Generation Using Wasserstein Autoencoders with a Mixture of Gaussian Prior
• Aggressive Language Detection with Joint Text Normalization via Adversarial Multi-task Learning