Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank

Semantic word spaces have been very useful but cannot express the meaning of longer phrases in a principled way. Further progress towards understanding compositionality in tasks such as sentiment detection requires richer supervised training and evaluation resources and more powerful models of composition. To remedy this, we introduce a Sentiment Treebank. It includes fine grained sentiment labels for 215,154 phrases in the parse trees of 11,855 sentences and presents new challenges for sentiment compositionality. To address them, we introduce the Recursive Neural Tensor Network. When trained on the new treebank, this model outperforms all previous methods on several metrics. It pushes the state of the art in single sentence positive/negative classification from 80% up to 85.4%. The accuracy of predicting fine-grained sentiment labels for all phrases reaches 80.7%, an improvement of 9.7% over bag of features baselines. Lastly, it is the only model that can accurately capture the effects of negation and its scope at various tree levels for both positive and negative phrases.

• Publication year

  2013

• Venue

  Conference on Empirical Methods in Natural Language Processing

• Publication date

  2013-10-01

• Fields of study

  Computer Science

• Identifiers
  DOI 10.18653/v1/d13-1170
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Multi-Step Regression Learning for Compositional Distributional Semantics

  2013cited by this paper
• A System for Real-time Twitter Sentiment Analysis of 2012 U.S. Presidential Election Cycle

  2012cited by this paper
• Large Vocabulary Speech Recognition Using Deep Tensor Neural Networks

  2012cited by this paper
• Improving Word Representations via Global Context and Multiple Word Prototypes

  2012cited by this paper
• A latent factor model for highly multi-relational data

  2012cited by this paper
• Semantic Compositionality through Recursive Matrix-Vector Spaces

  2012influential reference
• Compositional Matrix-Space Models for Sentiment Analysis

  2011cited by this paper
• Adaptive Subgradient Methods for Online Learning and Stochastic Optimization

  2011cited by this paper
• From machine learning to machine reasoning

  2011cited by this paper
• Semi-Supervised Recursive Autoencoders for Predicting Sentiment Distributions

  2011influential reference
• Experimental Support for a Categorical Compositional Distributional Model of Meaning

  2011cited by this paper
• Parsing Natural Scenes and Natural Language with Recursive Neural Networks

  2011influential reference
• Learning Continuous Phrase Representations and Syntactic Parsing with Recursive Neural Networks

  2010influential reference
• Dependency Tree-based Sentiment Classification using CRFs with Hidden Variables

  2010cited by this paper
• Factored 3-Way Restricted Boltzmann Machines For Modeling Natural Images

  2010cited by this paper
• Estimating Linear Models for Compositional Distributional Semantics

  2010cited by this paper
• From Frequency to Meaning: Vector Space Models of Semantics

  2010cited by this paper
• United we Stand: Improving Sentiment Analysis by Joining Machine Learning and Rule Based Methods

  2010cited by this paper
• Compositional Matrix-Space Models of Language

  2010cited by this paper
• Composition in Distributional Models of Semantics

  2010cited by this paper
• Distributional Memory: A General Framework for Corpus-Based Semantics

  2010cited by this paper
• Modelling Relational Data using Bayesian Clustered Tensor Factorization

  2009cited by this paper
• Multi-entity Sentiment Scoring

  2009cited by this paper
• Semantic Vector Products: Some Initial Investigations

  2008cited by this paper
• A unified architecture for natural language processing: deep neural networks with multitask learning

  2008cited by this paper
• A Structured Vector Space Model for Word Meaning in Context

  2008cited by this paper
• Combining Symbolic and Distributional Models of Meaning

  2007cited by this paper
• Dependency-Based Construction of Semantic Space Models

  2007cited by this paper
• Sentiment Composition

  2007cited by this paper
• Multiple Aspect Ranking Using the Good Grief Algorithm

  2007cited by this paper
• Contextual Valence Shifters

  2006cited by this paper
• Learning to Map Sentences to Logical Form: Structured Classification with Probabilistic Categorial Grammars

  2005cited by this paper
• Seeing Stars: Exploiting Class Relationships for Sentiment Categorization with Respect to Rating Scales

  2005cited by this paper
• A Neural Probabilistic Language Model

  2003cited by this paper
• Accurate Unlexicalized Parsing

  2003cited by this paper
• Minimizers, Maximizers and the Rhetoric of Scalar Reasoning

  2001cited by this paper
• Information, relevance, and social decisionmaking: some principles and results of decision-theoretic semantics

  1999cited by this paper
• Holographic reduced representations

  1995cited by this paper
• A natural history of negation

  1991cited by this paper
• Mapping Part-Whole Hierarchies into Connectionist Networks

  1990cited by this paper
• Recursive Distributed Representations

  1990cited by this paper
• Denial and contrast: A relevance theoretic analysis ofbut

  1989cited by this paper

• SEW: Strengthening Robustness of Black-box DNN Watermarking via Specificity Enhancement

  2026cites this paper
• Towards Privacy-Preserving LLM Inference via Collaborative Obfuscation (Technical Report)

  2026cites this paper
• LIME-LLM: Probing Models with Fluent Counterfactuals, Not Broken Text

  2026cites this paper
• Toward Ultra-Long-Horizon Sequential Model Editing

  2026cites this paper
• Zero-Order Optimization for LLM Fine-Tuning via Learnable Direction Sampling

  2026cites this paper
• Antibody: Strengthening Defense Against Harmful Fine-Tuning for Large Language Models via Attenuating Harmful Gradient Influence

  2026influential citation
• Prior-Informed Zeroth-Order Optimization with Adaptive Direction Alignment for Memory-Efficient LLM Fine-Tuning

  2026cites this paper
• Learning to Explain: Supervised Token Attribution from Transformer Attention Patterns

  2026cites this paper
• ICL-EVADER: Zero-Query Black-Box Evasion Attacks on In-Context Learning and Their Defenses

  2026cites this paper
• AQER: a scalable and efficient data loader for digital quantum computers

  2026cites this paper
• Hacking intelligence: Mapping the anatomy of adversarial threats in artificial intelligence with MITRE ATLAS

  2026cites this paper
• A unified optimization framework for backdoor attacks in large language models

  2026cites this paper
• Powering Up Zeroth-Order Training via Subspace Gradient Orthogonalization

  2026cites this paper
• Wave-Attractor-Tree: A Hierarchical Binary Tree Reduction Architecture for Efficient Sequence Modeling

  2026cites this paper
• Backdoor defense for large language models with weak-to-strong knowledge distillation

  2026cites this paper
• Quantifying the Effect of Test Set Contamination on Generative Evaluations

  2026cites this paper
• FiLoRA: Parameter-Efficient Fine-Tuning With Fisher Information-Guided Low-Rank Adaptation

  2026cites this paper
• Reasoning Stabilization Point: A Training-Time Signal for Stable Evidence and Shortcut Reliance

  2026cites this paper
• Curate-Train-Refine: A Closed-Loop Agentic Framework for Zero Shot Classification

  2026influential citation
• LLM-Driven Synthetic Text Generation for Privacy-Preserving Federated Learning

  2026cites this paper
• Demystifying Mergeability: Interpretable Properties to Predict Model Merging Success

  2026cites this paper
• RAPTOR: Ridge-Adaptive Logistic Probes

  2026cites this paper
• Dependable Artificial Intelligence with Reliability and Security (DAIReS): A Unified Syndrome Decoding Approach for Hallucination and Backdoor Trigger Detection

  2026cites this paper
• Training-free layer selection for partial fine-tuning of language models

  2026cites this paper
• Hidden Licensing Risks in the LLMware Ecosystem

  2026cites this paper
• Domain-Adaptive Multitask BERT with Graph Context Modeling for Code-Mixed Hinglish Sentiment Classification

  2026cites this paper
• Complex-Valued Unitary Representations as Classification Heads for Improved Uncertainty Quantification in Deep Neural Networks

  2026cites this paper
• The Cascade Equivalence Hypothesis: When Do Speech LLMs Behave Like ASR$\rightarrow$LLM Pipelines?

  2026cites this paper
• CF-STAR: Highly compressible adapters for model merging via centralized task vectors

  2026cites this paper
• Fusing Representation Spaces: A Projected-Fusion Approach to Contrastive Learning

  2026cites this paper
• LBKD: Rethinking Federated Backdoors for Low-Altitude Economy via LLMs and Bidirectional Knowledge Distillation

  2026cites this paper
• Semantic Text Kernels: A hybrid interpretable framework for deep semantic analysis in textual data

  2026cites this paper
• T3C: Test-Time Tensor Compression with Consistency Guarantees

  2026cites this paper
• Large-Scale Aspect-Based Sentiment Analysis with Reasoning-Infused LLMs

  2026cites this paper
• MI-PRUN: Optimize Large Language Model Pruning via Mutual Information

  2026cites this paper
• Enhancing Sentiment Classification and Irony Detection in Large Language Models through Advanced Prompt Engineering Techniques

  2026cites this paper
• Spectral Characterization and Mitigation of Sequential Knowledge Editing Collapse

  2026cites this paper
• Retrieval augmentation for out-of-distribution robustness in non-knowledge intensive in-context learning

  2026cites this paper
• A deep sentiment model combining ALBERT-driven context and EHO-optimized architecture

  2026cites this paper
• Beyond Hard Writes and Rigid Preservation: Soft Recursive Least-Squares for Lifelong LLM Editing

  2026cites this paper
• RouteMoA: Dynamic Routing without Pre-Inference Boosts Efficient Mixture-of-Agents

  2026cites this paper
• Persona Prompting as a Lens on LLM Social Reasoning

  2026influential citation
• Enhancing Public Healthcare Through VADER Sentiment Analysis: A Case Study on Patient Complaints

  2026cites this paper
• OSNIP: Breaking the Privacy-Utility-Efficiency Trilemma in LLM Inference via Obfuscated Semantic Null Space

  2026cites this paper
• Preserve-Then-Quantize: Balancing Rank Budgets for Quantization Error Reconstruction in LLMs

  2026cites this paper
• In-Run Data Shapley for Adam Optimizer

  2026cites this paper
• Revisiting Prompt Sensitivity in Large Language Models for Text Classification: The Role of Prompt Underspecification

  2026cites this paper
• Optimal Turkish Subword Strategies at Scale: Systematic Evaluation of Data, Vocabulary, Morphology Interplay

  2026cites this paper
• Learning a Generative Meta-Model of LLM Activations

  2026cites this paper
• Fine-Grained Model Merging via Modular Expert Recombination

  2026cites this paper
• Linearization Explains Fine-Tuning in Large Language Models

  2026cites this paper
• Context-Aware Counterfactual Data Augmentation for Gender Bias Mitigation in Language Models

  2026cites this paper
• Weight Decay Improves Language Model Plasticity

  2026cites this paper
• Learning fair representation for fine-tuning pre-trained language models.

  2026cites this paper
• ROAST: Rollout-based On-distribution Activation Steering Technique

  2026cites this paper
• Generalized and group spherical linear interpolation for token-level context compression.

  2026cites this paper
• Avey-B

  2026cites this paper
• Improving out-of-distribution detection in normalizing flows with synthetic outliers

  2026cites this paper
• SeedFlood: A Step Toward Scalable Decentralized Training of LLMs

  2026cites this paper
• Security and Privacy in LLMs: A Comprehensive Survey of Threats and Mitigation Strategies

  2026cites this paper
• Stop-Think-AutoRegress: Language Modeling with Latent Diffusion Planning

  2026cites this paper
• Model Merging in the Essential Subspace

  2026cites this paper
• Latent Variable Modeling for Controllable and Diverse Generation From Large Language Models

  2026cites this paper
• Entropy-Based Data Selection for Language Models

  2026cites this paper
• EEformer: Early Exiting for Transformer With Global-Local Exits and Progressive Fine-Tuning

  2026cites this paper
• Configurational patterns for forecasting customer satisfaction enhancement based on online reviews: A multi-attribute attitude perspective

  2026cites this paper
• BERT-JEPA: Reorganizing CLS Embeddings for Language-Invariant Semantics

  2026cites this paper
• The Alchemy of Thought: Understanding In-Context Learning Through Supervised Classification

  2026cites this paper
• Differential Privacy for Transformer Embeddings of Text with Nonparametric Variational Information Bottleneck

  2026influential citation
• Grad-ELLM: Gradient-based Explanations for Decoder-only LLMs

  2026cites this paper
• Nexus scissor: enhance open-access language model safety by connection pruning

  2026cites this paper
• Why LoRA Fails to Forget: Regularized Low-Rank Adaptation Against Backdoors in Language Models

  2026influential citation
• Sentiment Analysis on Movie Reviews: A Deep Dive into Modern Techniques and Open Challenges

  2026influential citation
• Q-realign: Piggybacking Realignment on Quantization for Safe and Efficient LLM Deployment

  2026influential citation
• EvasionBench: A Large-Scale Benchmark for Detecting Managerial Evasion in Earnings Call Q&A

  2026cites this paper
• CAPE: Generalized Convergence Prediction Across Architectures Without Full Training

  2026cites this paper
• HOSL: Hybrid-Order Split Learning for Memory-Constrained Edge Training

  2026cites this paper
• Differentially Private Subspace Fine-Tuning for Large Language Models

  2026influential citation
• A BERTology View of LLM Orchestrations: Token- and Layer-Selective Probes for Efficient Single-Pass Classification

  2026influential citation
• ContiguousKV: Accelerating LLM Prefill with Granularity-Aligned KV Cache Management

  2026cites this paper
• Robustness of Mixtures of Experts to Feature Noise

  2026cites this paper
• Abusive music and song transformation using GenAI and LLMs

  2026cites this paper
• AGZO: Activation-Guided Zeroth-Order Optimization for LLM Fine-Tuning

  2026cites this paper
• Mechanistic Analysis of Catastrophic Forgetting in Large Language Models During Continual Fine-tuning

  2026cites this paper
• Capsule-enhanced RoBERTa for hierarchical sentiment analysis on social media texts

  2026cites this paper
• LLM-VA: Resolving the Jailbreak-Overrefusal Trade-off via Vector Alignment

  2026influential citation
• Making Foundation Models Probabilistic via Singular Value Ensembles

  2026cites this paper
• Understanding Model Merging: A Unified Generalization Framework for Heterogeneous Experts

  2026cites this paper
• Exploring Knowledge Filtering for Retrieval-Augmented Question Answering

  2026cites this paper
• FlexLoRA: Entropy-Guided Flexible Low-Rank Adaptation

  2026cites this paper
• Quantifying Model Uniqueness in Heterogeneous AI Ecosystems

  2026cites this paper
• DimABSA: Building Multilingual and Multidomain Datasets for Dimensional Aspect-Based Sentiment Analysis

  2026cites this paper
• Beyond Local Edits: Embedding-Virtualized Knowledge for Broader Evaluation and Preservation of Model Editing

  2026cites this paper
• On the Relationship Between Representation Geometry and Generalization in Deep Neural Networks

  2026cites this paper
• Identification Techniques in the Internet of Things: Survey, Taxonomy and Research Frontier

  2026cites this paper
• UAT-LITE: Inference-Time Uncertainty-Aware Attention for Pretrained Transformers

  2026cites this paper
• BadTemplate: A Training-Free Backdoor Attack via Chat Template Against Large Language Models

  2026cites this paper
• When Shared Knowledge Hurts: Spectral Over-Accumulation in Model Merging

  2026influential citation
• Cost-Aware Model Selection for Text Classification: Multi-Objective Trade-offs Between Fine-Tuned Encoders and LLM Prompting in Production

  2026influential citation
• TokenCom: Vision-Language Model for Multimodal and Multitask Token Communications

  2026cites this paper