CMU-MOSI

The CMU-MOSI dataset is a benchmark that contains multimodal data for sentiment analysis, specifically designed to evaluate the performance of models in understanding sentiment from various modalities under challenging conditions.

Papers using CMU-MOSI (2)

FedUAF: Uncertainty-Aware Fusion with Reliability-Guided Aggregation for Multimodal Federated Sentiment Analysis2026Unimodal Training-Multimodal Prediction: Cross-modal Federated Learning with Hierarchical Aggregation2023 · 5 cites