Deniz's Reflection

Week 1:

I'm excited about the course! I'm quite interested in psychology, and I heard one of the instructors majored in Cognitive Psychology at Leiden University, that's so cool! It was also interesting to discuss how robots should complement rather than replace humans.
 

Week 2:

We started shaping our idea. At first, we thought of building a robot to help with pill dispensing, but value-sensitive design made us think about what people actually need. We realized that companionship and emotional support might matter more than functionality in this context. That shift made us rethink the project’s purpose, which I think was an important turning point.

Week 3:

We are pivoting towards a dancing companion robot for dementia patients. I find it a great idea to help get the PwDs some physical activity via dancing! We'll probably won't get to test this large scale, with actual dementia patients (I'm assuming that's outside of this course's scope), but I'd really like to observe the effects of such a system, applied in a real scenario.

Week 4:

We mainly worked on xwiki this week. I find the platform quite clunky; but it is a great system to present a report.

Week 5:

We had a presentation this week. Alex and Aleks did the slides, while me and Vladimir presented. it was nice to present, as I got to familiarize myself with all parts of our project, not just the ones that I worked on.

Week 6:

We decided on our prototype. We will be working with the Miro-E robot dog! This will be new for me. I looked up on how to connect to and work on actual robots.

Week 7: 

We met up with the team to try to work on Miro. It was a thrilling challange, and quite an experience; although we had mixed results. While we were able to connect to it, and make it move, and create sounds (only sine-waves at selected pitches though), we found quickly that 1) it was really clunky to work with, and 2) technology is quite outdated. so outdated that we found out that we couldn't send LLMs or implement a TTS inside Miro (Miro has Python 3.5 and OpenAI's API only supports Python 3.8 or higher). We instead resorted on implementing a conversational agent that has a listening module, speech module, and a Spotify API to play music (I was so glad that I took the Conversational Agents course last year, as I applied the knowledge I received in that course (and reusing some code from there) to implement this agent, Which we used to conduct the experiments with the participants.)

Overall, I really would have successfully liked to actually work with Miro. I feel a bit disappointed that we weren't able to, but it was also really valuable to gain experience in constructing a digital agent in place of Miro. It was still fun to interact with the robot, and I did feel as if it would work as intended in a real life scenario as well; but I would have liked to have tested it with a moving and dancing miro, since that was the original idea. Perhaps if I were to do this course all over again, I'd research how Miro works beforehand, and plan accordingly. We were kinda confused with the direction we wanted to take as a team, and we sort-of did it as we go; and we did not know that Miro was not capable of what we wanted to do with our project. I don't think we can say much from the results. I really would have liked to have a larger-scope course (maybe a 2-part course over 2 quarters or something) or a proper research project that would let me observe the effects our agent on a larger, more statistically significant population. I also really enjoyed working with my team. I feel like we got along very well, and we very often seemed to be on the same page, and we worked really efficiently. It made the whole course way more enjoyable.