Clone of BCI using Hierarchical Temporal Memory(HTM) or Deep Learning(DL)
The aim of our research is to implement Artificial Intelligence techniques to portable devices that measure electroencephalogram (EEG) signals. For this we use the Emotiv Epoch EEG device to classify imaginary and body movements of the body. Techniques such as Hierarchical Temporal Memory and Deep Learning are current being researched for interpreting EEG signals as they have shown to have great potential.
HTM and DL have shown to predict 4 imaginary and body movements with high accuracy. However, these algorithms have not been yet tested for higher number of classes such as the prediction movements of a whole arm or finger movements. In addition, most of the Learning has been done offline. Therfore, another possibility is to move towards online learning which will allow us to move to other fields of application such as controlling a phone or flying a drone with our mind.
- You must have taken or willing to take 'Machine learning' course.
- Basic knowledge on probabilistic models such as Bayesian Networks.
- It is not required to have previous knowledge about Brains or EEG signals. However, is a plus if you do.
Felipe Gomez; email: fegomezm[at]vub.ac.be or visit me at 10G.711