Investigating the effectiveness of attention mechanisms for EEG emotion recognition on DEAP dataset | |
| Author | Ravula, Greeshma |
| Call Number | AIT RSPR no.IM-21-08 |
| Subject(s) | Electroencephalography Pattern recognition systems Emotions--Computer simulation |
| Note | A research submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Information Management |
| Publisher | Asian Institute of Technology |
| Abstract | So far, many models have been used for recognizing emotions using EEG signals, ranging from traditional machine learning models to advanced deep learning models. However, at tention based models have not been fully investigated for EEG emotion recognition. In this study, we investigate the effectiveness of attention mechanisms. Specifically, we compare some variants with attention versus without attention on top of Bi-LSTM to evaluate the impact of attention mechanism on final outcome. We experiment our work on the DEAP dataset, which is benchmark dataset for EEG emotion recognition. Bidirectional Long Short Term Memory (LSTM) is used as a baseline model in our study. We have compared five variants of attention models incorporating attention on top of Bi-LSTM without attention. Bi-LSTM without attention, soft attention, self attention, multihead attention, multilevel attention, hard attention have been compared and achieved accuracies of 58.43%, 80.7%, 85%, 53%, 85%, 86% respectively. With this work, we have shed light over some variants of attention. Moreover, most of the variants perform well indicating that attention improves performance of a model and is a promising strategy for EEG emotion recognition. |
| Year | 2021 |
| Type | Research Study Project Report (RSPR) |
| School | School of Engineering and Technology |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Information Management (IM) |
| Chairperson(s) | Chaklam Silpasuwanchai |
| Examination Committee(s) | Dailey, Matthew N.;Attaphongse Taparugssanagorn |
| Scholarship Donor(s) | AIT Fellowship |
| Degree | Research studies project report (M. Eng.) - Asian Institute of Technology, 2021 |