Ednex

Inspiration

We were concerned about the difficulty of finding useful information and resources to complete projects, and found that many such sources were also not tailored to fit our current understandings. Thus, we were inspired to create Ednex, a platform that allows users to describe their prior knowledge and experience, as well as what they hope to learn, and use that information to find reputable sources tailored to their needs.

What it does

Ednex aggregates reputable sources from a multitude of locations and tags them effectively by using machine learning to analyze and categorize each one. After tags are assigned to each source, the text prompts provided by the user allow for the machine learning model to match up sources that best fit the user's request. Doing so is an iterative process; the user is able to adjust their own preferences and understandings to suit their needs and find better, personalized sources.

How we built it

The core of the Ednex UI is built upon Flutter for its adaptability to web, mobile, and desktop applications. Ednex is hosted on a Google virtual machine, part of Google Cloud, that is connected to a .tech url. A proof of concept exists for tag identification in sources such as articles and papers that is not yet implemented in the full release of Ednex yet, and will be fully integrated when tag identification and categorization is consistent and reliable.

Challenges we ran into

Coming up with a way to implement the machine learning models in a limited amount of time was a hard challenge to overcome. In order to create a machine learning model that would be capable of proper tag detection, we first started by researching for models that were similar to our own use cases. By doing so, we were able to find a good starting point for our own machine learning model. We were able to build upon an existing model API and implement it to test our own use cases for tag detection, and were ultimately able to end up with an adequate result for a proof of concept test for our machine learning implementation.

Accomplishments that we're proud of

Given the amount of experience that we had entering the project, and the fact that this was our first hackathon, we are proud of the finished product that we have prepared.

What we learned

Some of the most important things that we learned throughout the process of creating this project are the ability to take a concept and idea and bring it into fruition. We also learned that integration between different components of a project is nontrivial, and can become quite complicated if not tackled properly.

What's next for Ednex

We plan to invest more time into learning about machine learning techniques and resources that would be applicable to our project and, in the future, deploy a fully working version that utilizes these techniques to provide users with an experience.