Search

Rumor detection aims to detect rumors in a timely manner to prevent malicious rumors from misleading the public and disrupting social order. However, rumor detection suffers from the problem of imbalanced data. Existing methods of text generation and imbalanced learning are insufficient in addressing this imbalance because they are not specialized in rumor tasks. We propose a knowledge graph-based rumor data augmentation method: Graph Embedding-based Rumor Data Augmentation (GERDA), which simulates the generation process of rumor from the perspective of knowledge. To model the generation process of false information, we introduce knowledge representation in the process of text generation. To better learn the graph structured rumor data, we propose a graph-based rumor text generative model G2S-AT-GAN, which uses an attention-based graph convolutional neural network and a generative adversarial network for rumor text generation. Experiments show that our method is able to generate high-quality rumors of diverse topics and the generated rumors can further address rumor data imbalance for better performance in rumor detection. [ABSTRACT FROM AUTHOR]

Representation learning is a crucial and challenging task within multimodal sentiment analysis. Effective multimodal sentiment representations contain two key aspects: consistency and difference. However, the state-of-the-art multimodal sentiment analysis approaches failed to capture the difference and consistency of sentiment information across diverse modalities. To address the multimodal sentiment representation problem, we propose an autoencoder-based self-supervised learning framework. In the pre-training stage, an autoencoder is designed for each modality, leveraging unlabeled data to learn richer sentiment representations for each modality through sample reconstruction and modality consistency detection tasks. In the fine-tuning stage, the pre-trained autoencoder is injected into MulT (AE-MT) and enhance the model's ability to extract deep sentiment information by incorporating a contrastive learning auxiliary task. Our experiments on the popular Chinese sentiment analysis benchmark (CH-SIMS v2.0) and English sentiment analysis benchmark (MOSEI) demonstrate significant gains over baseline models. [ABSTRACT FROM AUTHOR]