MOSI

The MOSI dataset is a benchmark for Multimodal Sentiment Analysis that contains video clips annotated with sentiment labels, used to evaluate the effectiveness of models in understanding and interpreting human sentiment across multiple modalities.

Papers using MOSI (2)

Mixture of Disentangled Experts with Missing Modalities for Robust Multimodal Sentiment Analysis2026Learning Modality-Specific Representations with Self-Supervised Multi-Task Learning for Multimodal Sentiment Analysis2021 · 667 cites