HiBorn4/TensorFusion_Network_for_Multimodal_sentiment_analysis

This repository implements the Tensor Fusion Network (TFN) for multimodal sentiment analysis using the CMU-MOSI dataset. TFN integrates language, visual, and acoustic modalities to predict sentiment intensity, enhancing sentiment prediction accuracy by modeling unimodal, bimodal, and trimodal interactions.

21
/ 100
Experimental

The implementation uses LSTM-based modality embedding subnetworks for language (GloVe vectors), visual (FACET/OpenFace facial features), and acoustic (COVAREP) streams, which feed into a tensor fusion layer that explicitly computes three-fold Cartesian products to capture unimodal, bimodal, and trimodal feature interactions. The architecture culminates in a fully connected inference network supporting binary/five-class classification and regression tasks on the 2,199-utterance CMU-MOSI dataset, with ablation studies demonstrating that trimodal dynamics modeling is critical for performance gains.

No commits in the last 6 months.

No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 5 / 25
Maturity 1 / 25
Community 15 / 25

How are scores calculated?

Stars

13

Forks

5

Language

Jupyter Notebook

License

Last pushed

May 21, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/HiBorn4/TensorFusion_Network_for_Multimodal_sentiment_analysis"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.