Filtering by
- All Subjects: Technology
- All Subjects: Music
The purpose of this project is to create a useful tool for musicians that utilizes the harmonic content of their playing to recommend new, relevant chords to play. This is done by training various Long Short-Term Memory (LSTM) Recurrent Neural Networks (RNNs) on the lead sheets of 100 different jazz standards. A total of 200 unique datasets were produced and tested, resulting in the prediction of nearly 51 million chords. A note-prediction accuracy of 82.1% and a chord-prediction accuracy of 34.5% were achieved across all datasets. Methods of data representation that were rooted in valid music theory frameworks were found to increase the efficacy of harmonic prediction by up to 6%. Optimal LSTM input sizes were also determined for each method of data representation.
There are quite a few readily available products that one can buy if one looks past some of their flaws. A lot of these alarms either require a user to carry an extra communication device, or they are too big or expensive. The proposed solution merges all desirable features of a bike alarm into one module. In light of this, surveys were conducted to ascertain what these qualities would need to be. The top considerations for purchasing this alarm were how costly it would be, the false detection rate, and also the battery life. Additionally, the features that were most requested was the inclusion of a GPS and a camera. In order to incorporate these features, a three year plan was formulated which would culminate into a bike network in which each bike could communicate with other bikes. This would allow for an IOT network to be established, thus far exceeding expectations. The price point for this alarm is USD $10.00-15.00 and can come in a variety of colors. Additionally, this concept can be applied to many different scenarios, from protecting boats/jet skis and other aquatic vehicles, to houses as well. Furthermore, one could miniaturize this technology to be used in jewelry or accessories.