Inspiration

The inspiration for LOIE comes from a movie called "Her", which talks about a romantic relationship between an advanced AI Chatbot Samantha and a human writer Theodore. It's quite inspiring to see how a human and an AI Chatbot develop a deep relationship in the movie, which leads to inspiration for a virtual companion for people to combat loneliness, especially for young adults (18-24) who are worst affected during the pandemic. We found that the problem of loneliness carries a stigma that people tend to feel self-defeating to reach out for help, so they often hide their true feelings. Also, the mental health of young people is often neglected and they are left scrambling to figure out how to adjust to this new norm with very little societal support. This leads to the ideation of LOIE (product name inspiration: turn "lone" into "love" with "LOIE").

What it does

LOIE creates a safe and inclusive environment that allows young people to share their true feelings; and with users' input, it provides the tools they need to deal with loneliness. LOIE's main purpose is to encourage users to 1) connect with themselves or 2) connect with others when they feel lonely. Users can share anything with LOIE’s AI chatbot “Evie” whenever and wherever; in turn, Evie serves as a smart virtual friend to provide users with emotional support and guidance using technology like Deep Neural Learning, Natural Language Processing and Emotion AI. Furthermore, LOIE uses Machine Learning based on users' data to connect users with like-minded people through online communities, forming support groups and help each other out.

How we built it

We built the application using python and deep learning. Libraries such as nltk, keras, Tensorflow and NumPy were used to create the model and the chatbot. The app is built in three parts the first one is Intents file (datasets) which includes all the emotions of the humans, what questions humans usually ask and how the bot responds. Then we processed the data for training the model so that Evie can find the most suitable patterns so that she can respond to the users appropriately. We utilized Tkinter for the application GUI implementation. We developed a simple interface of the BOT which includes all the buttons and text widgets.

Challenges we ran into

We faced technical challenges during development since most of our members were new to Deep Learning. Using specific packages such as nltk and TensorFlow, fixing coding and file running errors were the largest challenges.
Working with people living in different places around the globe was not easy as each member’s timezone is different. Writing an Algorithm to find the right pattern for the Response was one of the big challenges we faced.

Accomplishments that we're proud of

We are very proud of completing a project based on deep learning which was new to most of us. We developed the interface for LOIE which is extremely user-friendly and the algorithm which is responsible for finding the best possible response.

What we learned

Researching mental health issues in young adults during the pandemic helped us realize how serious it is and that we need to take action. We were able to acquire new technical knowledge in different python libraries and product design. It was also a great opportunity to learn how to work on a project as a team remotely. Collaboration between team members with different skills is vital to the flow and development of this project - without frequent voice calls there would have been a miscommunication.

What’s next for LOIE

Some additional accessibility features can be added for a better user experience. We could also add more community features so that the users can connect with other users. GUI will be more user-friendly. Large amounts of datasets would be added. LOIE would be more secure as we would be Using Facial Recognition to unlock LOIE and Speech Recognition would soon be available. Soon, Evie will be available on Discord with many cool features like play music, send memes and more.

Technical References

Tensorflow : https://github.com/tensorflow/tensorflow NLTK : https://www.nltk.org/ NUMPY:https://numpy.org/doc/ PANDAS: https://pandas.pydata.org/docs/ TKINTER: https://docs.python.org/3/library/tkinter.html

Research References

Beresin, G. (n.d.). Why Are Teens So Lonely and What Can They Do to Combat Loneliness? The Clay Center for Young Healthy Minds. https://www.mghclaycenter.org/parenting-concerns/teenagers/why-are-teens-so-lonely-and-what-can-they-do-to-combat-loneliness/ CBC News. (2021, January 30). Solitude and loneliness wearing down even the most connected generation. https://www.cbc.ca/news/canada/montreal/mental-health-youth-covid-19-quebec-isolation-1.5893989 Statistics Canada. (2021, March 18). Survey on COVID-19 and Mental Health, September to December 2020. Statistics Canada. https://www150.statcan.gc.ca/n1/daily-quotidien/210318/dq210318a-eng.htm Walsh, C. (2021, February 17). Young adults hardest hit by loneliness during the pandemic. The Harvard Gazette. https://news.harvard.edu/gazette/story/2021/02/young-adults-teens-loneliness-mental-health-coronavirus-covid-pandemic/

Built With

Share this project:

Updates