CMU-MOSI and CMU-MOSEI Benchmark Splits for Multimodal Sentiment Analysis

A dataset split for benchmarking multimodal sentiment analysis models, published by Wenyan Xiao on 2026-05-12. It contains textual, acoustic, and visual information aligned for emotion interpretation. The data is stored in an XLS file with a size of 5.5 KB.

Use Cases

• Benchmarking multimodal fusion models based on textual, acoustic, and visual features.
• Training text-anchored alignment modules based on the described semantic anchoring framework.
• Evaluating model robustness against heterogeneous quality and weak correlation among nonverbal modalities.
• Conducting ablation studies on hierarchical gating and contrastive learning components.

Strengths

• Data is associated with established CMU-MOSI and CMU-MOSEI benchmarks.
• Released under a permissive CC-BY-4.0 license.
• Published on 2026-05-12, indicating recent availability.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

figshare

Freshness

Last updated 2026-05-12 17:38:25; freshness should be verified.