Abstract:In a real-world scenario, an object could contain multiple tags instead of a single categorical label. To this end, multi-label learning (MLL) emerged. In MLL, the feature distributions are long-tailed and the complex semantic label relation and the long-tailed training samples are the main challenges. Semi-supervised learning is a potential solution. While, existing methods are mainly designed for single class scenario while ignoring the latent label relations. In addition, they cannot well handle the distribution shift commonly existing across source and target domains. To this end, a Semi-supervised Dual Relation Learning (SDRL) framework for multi-label classification is proposed. SDRL utilizes a few labeled samples as well as large scale unlabeled samples in the training stage. It jointly explores the inter-instance feature-level relation and the intra-instance label-level relation even from the unlabeled samples. In our model, a dual-classifier structure is deployed to obtain domain invariant representations. The prediction results from the classifiers are further compared and the most confident predictions are extracted as pseudo labels. A trainable label relation tensor is designed to explicitly explore the pairwise latent label relations and refine the predicted labels. SDRL is able to effectively and efficiently explore the feature-label relation as well as the label-label relation knowledge without any extra semantic knowledge. We evaluated SDRL in general and zero-shot multi-label classification tasks and we concluded that SDRL is superior to other SOTA baselines. Furthermore, extensive ablation studies have been done which reveal the effectiveness of each component in our framework.

A Soft Label Strategy for Target-Level Sentiment Classification

Soft-label recover based label-specific features learning

Semi-supervised Target-Oriented Sentiment Classification

Semi-Supervised Dual Relation Learning for Multi-Label Classification

Roll With the Punches: Expansion and Shrinkage of Soft Label Selection for Semi-supervised Fine-Grained Learning

Effective Token Graph Modeling using a Novel Labeling Strategy for Structured Sentiment Analysis

Effective LSTMs for Target-Dependent Sentiment Classification

A Soft-label Method for Noise-tolerant Distantly Supervised Relation Extraction

Dynamic Auxiliary Soft Labels for Decoupled Learning

Learning From Biased Soft Labels

A multi-label classification based approach for sentiment classification

A Unified Labeling Model for Open-Domain Aspect-Based Sentiment Analysis

Learning label smoothing for text classification

SoftMCL: Soft Momentum Contrastive Learning for Fine-grained Sentiment-aware Pre-training

Semantic-Unit-Based Dilated Convolution for Multi-Label Text Classification.

Text multi-label sentiment analysis based on Bi-LSTM

A Soft Contrastive Learning-based Prompt Model for Few-shot Sentiment Analysis

Learning to Purify Noisy Labels Via Meta Soft Label Corrector.

Estimating Soft Labels for Out-of-Domain Intent Detection

Modeling Label Semantics for Predicting Emotional Reactions

Learning to Classify Open Intent via Soft Labeling and Manifold Mixup