Final Project - Week 0 (Feb. 16 - Feb. 23)

Goals for this week:

• Determine which project I will do

• Pitch said project to class on Monday 16 February

• If necessary, make changes/adapt project based on feedback

• Begin coding/software portion

Progress

• I decided on the Breaking and Entering robot to pitch because it seemed both useful and fun to make. Also, it seemed like the easiest robot to pitch out of my idea's which made it a great idea at the time. Unfortunately, I forgot the fact that there are legal issues in place that prevent me from creating this robot. Thus a new project had to be chosen. I decided upon the multi-language robot. (Monday 16 February, 2015)

• Talked to Shani (one of the T.A.s for the course) about her thoughts on my project. Shani recommended not doing the speech recognition portion of the design until the end, as she would not be able to help as much with this part if I got stuck. Also, she told me that she has not had experience with using a board other than the Arduino Uno ( I was considering the Intel Edison or Gallileo) , and so there would also be less help available for debugging on this. (Wednesday 18 February, 2015)

• My thoughts after discussion: I'm a little wary, but still excited to create this project. I've never worked in A.I. before and this class didn't really cover it, so this would be completely new ground for me. Takeaway is to start early and realize that support system might not be enough to help.

• Talked to Professor Heather Pon-Barry about the feasibility of this project. Professor Pon-Barry's research focus is on developing new technologies for speech interfaces, which I'll hopefully be able to use to get speech recognition by the robot--enabling conversation. Professor Pon-Barry said to start early, as this is an ambitious project, and that it is definitely feasible. She suggested using Sphynx or Kaldi to do the speech interface. (Thursday 18 February, 2015)

• My thoughts after discussion: Speech interface will definitely come last. First focus is to get text based version to work. Then I will add speech recognition.

• Created first version of Chatterbot in Java (Friday 19 February, 2015)

• Problems encountered:

• Case, accent and punctuation specific

• I think I will be able to figure out case, but punctuation and accent may be a bit more difficult

• Only one option for how to respond to a question. This makes it slightly boring to play around with right now.

• Meeting with Professor Audrey St. John to set-up Intel Galileo (Tuesday 23 February, 2015)

• Spent 2 hours trying to get Galileo to work. It was somewhat successful, but not completely. See Galileo Set-up Post for more in-depth info.

Next Steps:

• Write the Project Proposal and Design Proposal, which are due in class Monday, 23 February, 2015

• Create second version of Chatterbot by Wednesday, 25 February, 2015.

• Add more functionality to the program: ignore case, add more responses, etc...

• Try using a text to speech software library to have computer say their response (?)

• Get Intel Galileo to function ASAP!

Inspirational Pictures to Motivate:

#finalproject #weeklyupdate