Quantum Cognitively Motivated Decision Fusion for Video Sentiment Analysis

Authors: Dimitris Gkoumas, Qiuchi Li, Shahram Dehdashti, Massimo Melucci, Yijun Yu, Dawei Song827-835

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on two benchmarking datasets illustrate that our model significantly outperforms various existing decision level and a range of state-of-the-art content-level fusion approaches.
Researcher Affiliation Academia 1 The Open University, Milton Keynes, UK 2 University of Padua, Padua, Italy 3 Queensland University of Technology, Brisbane, Australia 4 Beijing Institute of Technology, Beijing, China
Pseudocode No The paper describes the methodology in text but does not include any formal pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any explicit statement about releasing its source code or a link to a code repository.
Open Datasets Yes We evaluated the proposed model on two benchmarking datasets, namely, CMU Multimodal Opinion-level Sentiment Intensity (CMU-MOSI) (Zadeh et al. 2016) and CMU Multimodal Opinion Sentiment and Emotion Intensity (CMU-MOSEI) (Bagher Zadeh et al. 2018).
Dataset Splits Yes We estimated the uni-modal observables from training plus validation sets, and then we used the learnt observables for predicting the utterance sentiment on the test set.
Hardware Specification No The paper does not provide specific details about the hardware used to run the experiments, such as GPU or CPU models.
Software Dependencies No The paper mentions using 'MATLAB fsolve function' and techniques like 'Bi-GRU layers' and 'softmax layer', but does not specify exact version numbers for any key software components or libraries.
Experiment Setup Yes We randomly initialized the parameters {θG, θL, θV , θA, φL, φV , φA} [0, 2π] for uni-modal observable estimation, and {θT , φT } [0, 2π], ηT [0, 1] for utterance state estimation. The random initialization was repeated for 200 times to obtain the optimum solutions by calculating the minimum sum of squared loss. We used Bi-GRU layers (Cho et al. 2014) with forward and backward state concatenation, followed by fully connected layers.