Memory Mate

Home screen. The button pulses to draw attention and indicate its functionality
Gathering personal info gradually for user ease, avoiding overwhelm
Chat page serves as the cornerstone of Memory Mate. It provides companionship and build a rich personality model
Personalized cognitive exercise, designed to stimulate various cognitive functions
This minimum score of result is set to 50 to prevent users from feeling disheartened and help sustain their interest and motivation

Inspiration

The number of elderly individuals in the population is increasing due to improvements in healthcare and declining fertility rates. Among this growing demographic, individuals with dementia face unique challenges that require specialized care and support. However, treating dementia is complex and costly, and access to necessary resources and support is limited. Existing cognitive exercising apps are ineffective in providing meaningful cognitive stimulation for individuals with dementia. It is crucial to develop innovative solutions that are accessible, engaging, and effective in promoting cognitive health among the elderly, particularly those with dementia.

What it does

Our solution is Memory Mate, an intuitive and user-friendly platform designed to provide tailored support for the elderly who may not be familiar with technologies. Our solution includes the development of an intuitive and user-friendly app specifically designed for the elderly, even those who may not be familiar with technologies.

We understand the significance of companionship and support in the lives of elderly individuals, particularly those with dementia. To address this, our solution incorporates a personalized and friendly chatbot feature. The chatbot serves as a companion and a source of consultation, providing reassurance and guidance whenever needed.

We also understand the importance of medical evaluation and personalized cognitive workouts. Based on these evaluations, personalized cognitive exercises are developed and tailored to the specific needs of each individual.

How we built it

The following is a comprehensive overview of the integration of Gemini AI into Memory Mate.

When users send an image along with a message to the chatbot, Gemini 1.0 Pro Vision steps in, specialized for image analysis. This innovative model translates the image into text, seamlessly integrating visual data into the conversation. This text is then incorporated into the final prompt for Gemini 1.0 Pro, enabling seamless communication with the users.

Gemini AI plays a pivotal role in evaluating users' dementia levels within Memory Mate. Memory Mate saves the chat history and the latest quiz results directly on the user's device. Periodically, this data is transmitted to Gemini AI for analysis. Gemini AI then examines the collected data, considering crucial factors such as behavioural and psychological symptoms.

Memory Mate then sends the evaluation data gathered and the chat history to Gemini AI. Leveraging this rich dataset, Gemini AI crafts multiple-choice questions specifically designed to challenge and engage the user. What sets these exercises apart is their adaptability. Gemini AI carefully selects questions with the appropriate mixture of difficulties, considering the user's dementia level. This ensures that the exercises are both stimulating and achievable, maximizing their effectiveness in promoting cognitive health. Furthermore, Gemini AI ensures the relevance of the questions by drawing from the user's chat history, making the exercises more meaningful and engaging.

Challenges we faced

We encountered several challenges during our project. We joined the hackathon at a late stage, which posed significant time constraints. Any unforeseen obstacles could potentially lead to failure. Additionally, we struggled to find just the right prompt for Gemini AI to make a personalized chatbot and cognitive exercises. Every word in the prompt could change the outcome, so we had to try lots of experiments in Google AI Studio to get it right.

What we learned

This project is our debut in mobile app creation. We discovered how simple it is to make a mobile app for both Android and iOS using React Native, thanks to our previous experience with the React framework.

We also found out that careful planning upfront can save us a ton of time later on. While sketching out our app with Figma, we realized some tweaks were needed to our initial plans. It made us realize how much trouble we could have faced if we discovered those issues during the actual development phase.

Lastly, we were amazed by how smart Gemini AI is. It consistently outperformed our expectations, showing us the incredible possibilities of using advanced language models to improve the world.