NeuroTechSC

11/2 Week Updates

Data Team

This week, the Data team was able to process the last data files from the last person who has the hardware. They are still waiting on the final data set from the current person who has the hardware. While the team is waiting, they will be working on the presentation for the competition.

Machine Learning

This week not much happened, the ML team is still waiting on new data to come. One thing that has been started is the ML team slides for the presentation which goes over what the team is doing and explains the model which uses convolution. Convolution in terms of ML is the mixing of information and can be used to simplify complex equations. The model consists of 3 convolution layers, temporal convolution learns the frequency filters, depthwise convolution learns spatial filters, and separable convolutions to summarize the feature maps and merges all three outputs together. This way analyzing the data will be easier because the information that is not needed is filtered out.

UI Team

This past week, the UI team had a more relaxed workload, focusing on final touches to the interface. As of today, the following photo is how the user-interface looks like. This week, the team will work on fixing edits on the interface, like changing the color scheme to have a more readable text as well as changing the prediction output from 1/0 to yes/no. The team will also change the countdown to start with the click of a button as opposed to starting automatically. The countdown is utilized by the user to know when to vocalize their response to the question. With this change, it will incorporate the hardware by letting it know when to start recording the response. Photo