search

kbmcgowan / Wildcat-Voices

CS 499 Senior Project for the University of Kentucky. The Wildcat Voices project aims to predict the location of a speaker, given short audio samples.
0 stars 0 forks source link

Tam Nguyen's Lab Notes #3

Open tam-kng opened 4 years ago

tam-kng commented 4 years ago

This issue is to log my personal notes during work on the Wildcat Voices project.

tam-kng commented 4 years ago

Team completed project plan documents. Personal contributions include:

Words written for the project plan total 689.

tam-kng commented 4 years ago

Today, our team revised the project plan documents. This was done to ensure completeness and provide some more context on our project. Specific aspects expanded upon include the Introduction's Product Overview in the Requirements Specification document, and personal role in the Planning and Estimating document.

The new word count written total for the project plan is 719.

tam-kng commented 4 years ago

Today, I wrote the first draft of my ethics essay. I chose the topic of user interfaces to discuss. The word count was 921 words.

tam-kng commented 4 years ago

Today, I revised my rough draft of my ethics essay using the notes my reviewer (Lucas Harten) made. The new total word count is again 921 words.

tam-kng commented 4 years ago

Today, our team met with our customers. We discussed storing our code in a public Github repository, and the licensing we would use (likely GPL). The code for the current Wildcat Voices is stored in the University of Kentucky web server.

Recently, it was discovered that the Wildcat Voices data was lost due to a bug in the code of the website. This bug was fixed; however, hundreds of voice recordings were unable to be saved. Therefore, we will be initially using the George Mason audio data to train our machine learning model. Although the George Mason audio data was recorded in a research lab, so as to limit echo from degrading the data, it is aged, so there may be differences in noise compared to audio samples recorded by modern smartphones. Along with the changes that we are making within the scope of this project, the customer will also be tweaking the website. Due to the aforementioned issue with the data, we will be attempting to collect more data after developing our Phase 2 of the project, the machine learning model. It's possible that we can ask our CS 499 class to provide some audio samples.

The machine learning model itself will be built to be modular so different methods can be used (SVM or neural nets). Once we develop our infrastructure for the assignment due next week, and move into development, we plan to test with the University of Kentucky web server.

tam-kng commented 4 years ago

Today, our team worked on the architecture document and architecture presentation assignments. I contributed to the discussion of the user interface of the Wildcat Voices website. Our team decided that since our primary focus is to develop a machine learning model and allow researchers to view its accuracy in guessing users' places of origin, the amount of user interface changes to the existing website would be kept to a minimum. This will be done by appending the system's guess, as well as a text box for users to fill, to the existing confirmation pages.

For researchers, an additional tab will also be created that will allow them to view system information upon logging in. This information includes how many submissions to the Wildcat Voices project there currently are as well as the accuracy of the system's guesses compared to the user-entered information of their place of origin.

A new issue with our project is that we have found that the current version of the Wildcat Voices website, does not function correctly in recording voice information. We have tried multiple browsers, including Chrome, Firefox, Safari, and Brave. Both desktop and mobile browsers were used. We plan to speak with our customers about this and determine the cause of this issue, as we have not begun the development phase of this project.

My total word count of the architecture assignment is 223 words. My total word count in the architecture presentation is 64 words.

tam-kng commented 4 years ago

Today, our team gave an update on our progress. I also filled out the senior survey and messaged Mr. Beck, and filled out the Google Forms survey. The combined total of both of these was 1180 words.

tam-kng commented 4 years ago

Today, I worked on the test plan for our Wildcat Voices project. This involved outlining the acceptance test cases for our user interface. In total, I wrote about 247 words.

tam-kng commented 4 years ago

Today, I reviewed our final presentation and added small notes that I will use to prepare for tomorrow's presentation. I will be turning in my developer notes, as well as my resume, to increase my word count as well.

I have enjoyed this class, and I believe keeping this log throughout the semester has helped me keep track of my work and progress.