Xplicate

We would like to be considered for GCP's challenge and Domain.com's challenge.

We're aiming to a create tool that pierces through the veil of misunderstanding that presents itself as fear during a time like the COVID 19 pandemic. We wanted to create something that helped an average person understand actual and reliable sources of data.

(xplicate.tech's c-name is not yet updated but it exists :-) )

Xplicate: analyze and develop (an idea or principle) in detail.

Our Inspiration:

There are researchers all over the world working hard to investigate the SARS-CoV-2 virus and the impact of this disease. Researching is a long process that can be broke down into 5 key steps: Locate and Define Issue or Problem, Creating a Research Plan, Collecting Data, Interpreting Research Data, and Report Research Findings

Our goal was to simplify the first step (locate and define problem) in order to make researching faster. The usual researching process can feel long because of the difficult language associated with problems and papers. However, if difficult words were defined as we could read it would make this researching process a little more faster.

Majority of the information the common person uses to find out more about COVID-19 is second hand information. With a way to make any person more quickly understand research papers, teens and adults will not get information that is misleading and understand the true severity. By finding information from a first hand source it will not only provide accurate details, but also allow the common man to learn that there are also multiple solutions occurring in this direction.

Essentially our product is made for two main demographics: researchers that want to find solutions for COVID-19 more quickly and mislead people who need more accurate information regarding COVID-19

What It Does:

Xplicate is a webapp with one main function. A user can upload a pdf regarding any topic that they are having a difficulty comprehending. This pdf will then be given to a user with a list of the most difficult words that are included and their definitions. With the meanings of the words right next to the paper the idea is to provide a much simpler way of reading. The user can make an account if they would like to save their pdf and definitions.

How It Does It:

We built Xplicate with GCP technologies. That is we used google cloud functions to host all of our serverless backend code; that code was written in typescript that was then transpiled to javascript and run in the Node.js runtime. There were four main API's that we used in order to accomplish Xplicate's task. We used the Google Cloud Storage API to store the users PDF's after they uploaded them to our service. This was then passed into the Google Vision API to run OCR on the PDF and convert the document to a plain text format. That plain text could then be passed to the the Google Language API which gave us a salience statistic which we combined with a few other measures to determine the best words to define and explain to the user. We then use a variety of contextual techniques as well as the Wikipedia API to get palatable definitions of difficult concepts.

Challenges We Ran Into:

This whole project was a couple of levels higher in complexity than what either of us had attempted before. We were trying something new with using Google Cloud and as a result ran into a lot of just basic configuration and up and running issues. We were upset that our cloud functions couldn't be completely finished by the end of 24 hours but we did get to learn about debugging Node.JS environments.

What We Learned

We learned that its very important to come up with a proof of concept of your final idea before completely diving into it. Even though our cloud functions aren't completely done, since we were able to make our proof concept we can show that our idea really works and we know the general direction we want to take it as we continue it. We learned that its a lot more important to learn about the issues that you run into rather than just brushing them off and ignoring them.

What's next for Xplicate:

We plan to continue to develop this project further! Eventually we would like to train the AI enough for it to be able to summarize difficult documents into simpler terms. Defining hard words was the first step in doing this, but Xplicate has an immense ability to grow.

Built With

• google-cloud
• google-cloud-firestore
• google-cloud-functions
• google-cloud-natural-language-processing
• google-cloud-platforms
• google-cloud-vision
• javascript
• optical-character-recognition
• python
• typescript