CMU-MOSI: Multimodal Sentiment Benchmark Performance Results

Experimental results comparing multimodal fusion methods on the CMU-MOSI sentiment benchmark. The dataset likely contains performance metrics from a federated learning framework that uses tensor decomposition for privacy-aware training. It was authored by Li Wan and uploaded to figshare on 2026-05-06.

Use Cases

• Benchmarking multimodal sentiment analysis algorithms based on the CMU-MOSI dataset mentioned in the description
• Evaluating federated learning frameworks for multimodal data fusion based on the privacy-aware decentralized training approach described
• Studying the effectiveness of tensor decomposition for modeling intermodal correlations based on the high-order tensor fusion method described

Strengths

• Dataset is openly licensed under CC-BY-4.0.
• Results are based on a recognized benchmark dataset (CMU-MOSI).
• The description details a specific methodological framework (federated learning with tensor Tucker decomposition).

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is very small (5.5 KB), suggesting limited scope.

Provenance

Source

figshare

Collection Method

Likely generated from experimental results of a federated learning framework.

Freshness

Last updated 2026-05-06 17:23:13; freshness should be verified.