Week #6 #
Presentation: #
https://drive.google.com/file/d/1dV_cJKX9vUy2w15olKWucj-4bLoTzUln/view?usp=sharing
This week we did the first iteration of final presentation. Some slides still need to be improved, and a few additional slides should be added to illustrate important details
Weekly Progress Report #
Our team did final steps for MVP version.
Backend & ML #
During this week, our team made significant strides in enhancing the emotion detection system and refining its functionalities to provide a more intuitive and accurate user experience. The focus was on iterating the existing model to improve its sensitivity and accuracy in detecting a wider range of emotions. This involved fine-tuning the algorithms and expanding the dataset to include more diverse emotional expressions, which enabled the system to better understand and respond to user inputs.
One of the key highlights of the week was the successful implementation of a new function that allows for the storage and transmission of user dialogues. This feature is crucial as it enables the system to maintain a record of conversations, which can be used for various purposes, such as improving the training data for the emotion detection model, providing users with a history of their interactions, and enabling seamless continuation of dialogues across different sessions and devices.
Design #
We also made a final tweak to our design to improve the overall user experience. We added a placeholder text, “I feel …”, to the input field where users interact with the service. This small but significant change serves as a request template, guiding users on how to communicate their emotions to the system effectively. This placeholder aims to make the interface more intuitive and user-friendly, helping users express their feelings more easily and ensuring that the system can respond appropriately.
Frontend #
Our team achieved significant milestones, with the full implementation of the web design standing out as a major accomplishment. We successfully uploaded the final working version of the web interface, marking a crucial step in our development process. This accomplishment is the result of meticulous planning, collaborative effort, and relentless dedication from our entire team.
The journey to this point involved a series of iterative design and development cycles. We began by meticulously crafting the user interface, ensuring that it was both aesthetically pleasing and functionally robust. Our design team worked tirelessly to create a seamless and intuitive user experience, paying close attention to every detail, from color schemes and typography to layout and navigation.
Challenges & Solutions #
One of the major hurdles we faced was configuring Google authentication for our web application. Initially, we ran into several issues that hampered the smooth integration of this feature. To address these issues, we turned to step-by-step guidance from Google’s developer documentation and various online resources. This approach proved to be highly effective.
On the ML side, we focused on setting up and managing interactions with a vector database, which is critical for handling the large-scale data operations required by our system. However, this task presented numerous challenges and turned out to be more time-consuming than anticipated.
The primary difficulty lay in optimizing the database to efficiently store and retrieve high-dimensional vectors. Our team spent considerable time experimenting with different configurations and conducting extensive testing. We also engaged in thorough code reviews and debugging sessions to ensure that the interactions between our ML models and the vector database were both efficient and reliable.
Conclusions & Next Steps #
In conclusion, this week was marked by significant advancements across all aspects of our project, including backend, design, and frontend development. Moving forward, our next steps will involve finalizing the remaining presentation slides.