StudyScape

Inspiration

One of my biggest struggles in learning has been memorizing concepts and vocabulary, especially with the rise of virtual learning and our increased reliance on technology. One effective method I discovered is storytelling. By weaving concepts into a narrative, I could remember the material better by recalling the story. This insight sparked the idea for this project of using generative AI to help people study using stories. We then proceeded to add to our vision for this tool with the goal of increasing retainment of information and personalizing the experience for the user.

What it does

StudyScape immerses users in a universe where they make decisions in an imaginative world that integrates the concepts they need to learn, allowing them to absorb information subconsciously. This project emphasizes the importance of continued personalization, by taking into account the mood of the user through every response, as well as incorporating regular quizzes that see whether the user retains information better with certain text style, genre, tone, and length of time and seamlessly incorporating these user analytics into the storytelling.

How we built it

We used Google Breadboard to create this project. We used the 'specialist' feature to perform certain functions in the app like deciphering user input, generating stories, generating questions for the user, and calculating study time allocated for the session. We used the 'deflator' feature and the 'output' feature to store and pass through information. Additionally, we used multiple loops to have conversations with our input mechanism which played the role of the student.

Challenges we ran into

We faced several challenges while developing the project. One issue we had was with organizing the connections of various blocks of functions in order to produce the sequential output desired. We also had difficulty understanding the loop feature and using it to sustain a complex conversation with the student using context. Another issue we had was that, as this is a new AI platform, finding online support was difficult, resulting in a lot of trial and error before we could get each feature to work properly.

Accomplishments that we're proud of

We are proud of how we were able to make the looper run through the storyteller and student to get a conversation flowing naturally. Achieving this required several beta versions before we reached out desired functionality. We are also proud of the computer's ability to take into account the various preferences of the user, and use it to adapt its output accordingly. This adaptability is a core aspect of our project, enabling it to understand and cater to the individual needs of the user.

What we learned

During the Hackathon, we were able to play with the Google Breadboard tool and understood its seemingly endless scope. Using Google Breadboard was a very different experience from regular programming because there was no need for direct input-to-output communication, we were able to give the computer a bunch of information and context, and it could use what it needed from that information to respond, which was very cool. Giving the computer a lot of context was really game-changing because it allowed the computer to really understand the user and personalize the experience in any way possible. It was also fascinating to see how the computer could be creative with the stories and the potential for more complex systems of generation such as individual roles for idea generation, multiple editors, and multiple writers, mimicking a real-life story-generating experience.

What's next for StudyScape

We would like to have the output mechanism of our project more nuanced by using the mood of the user input to not only change the plot of the story, as it exists now but also change the way it outputs the story. This feature would help the output mechanism match the mood of the user as well as be able to provide a supportive tone and reassurance when needed that a human study helper would give. We would also like to change the quiz feature of the project to be more conversational. In that, the quizzer should be able to have a conversation with the user about the question, providing context and hints if needed. We would like to include a conversation with the user at the beginning regarding the user's preferences for the story, whether the user would like a story based on real life to understand the real-world context of the issue, or whether they would like certain themes in their fictional story like fantasy, romance, science fiction, etc to provide a more enjoyable and engaging reading experience. While the project gets a lot of information from the user on how to react and respond in a personalized way, in the future, we would also like to partner with educators and storytellers to provide insight on how to make the output tool more effective and a better experience for the user.

Built With

• breadboard
• natural-language-processing