This is the code for Tensor Fusion Network for Multimodal Sentiment Analysis published in EMNLP 2017 and orally presented in multimodal session. The code is quite straight forward. Please download the CMU-MOSI dataset using CMU Multimodal Data SDK or my website. The data_loader.py helps you load the data in the correct format, however I suggest using the CMU Multimodal Data SDK for better loading as the directory structure of CMU-MOSI changes when we add new features. The code for the algorithm is in tf_mosi.py.
Please cite the following publication if you are using this code:
@inproceedings{tensoremnlp17,
title={Tensor Fusion Network for Multimodal Sentiment Analysis},
author={Zadeh, Amir and Chen, Minghai and Poria, Soujanya and Cambria, Erik and Morency, Louis-Philippe},
booktitle={Empirical Methods in Natural Language Processing, EMNLP},
year={2017}
}