2. pip install -r requirements.txt
3. python3 main.py
Note: Don't forget to add your own OAuth2 keys for GitHub, Google in .env in place of <GITHUB_CLIENT_ID>, <GITHUB_CLIENT_SECRET> and <GOOGLE_CLIENT_ID>, <GOOGLE_CLIENT_SECRET> respectively.
IntelliMed was a product of necessity, particularly because there's nothing quite like it. No product or service has come close to what IntelliMed has achieved- deducing complex medical reports to simple ones, readable by everyone. We noticed that medical reports are often complex and hard to comprehend and to our surprise, there wasn't a single solution that could make this easier.We are proud to announce that IntelliMed has done just that. IntelliMed uses one-of-a-kind Data Analysis and Machine Learning powered by the BERT technology (the same technology used by Google Search!) and NVIDIA's Cuda (the world's largest and best GPU computing platform) to summarize, annotate and analyze medical reports, making the reading of medical documents and papers easier for the patient. It also uses Google Cloud Services to jot down keywords and their meanings to further simplify documents. As soon as you enter IntelliMed, with its colorful UI and user-friendliness, you're bound to fall in love. There's simply nothing like it. We used BERT, a groundbreaking AI technology by Google (the same technology employed in the world's most popular and powerful Search Engine) to accurately summarize medical reports uploaded by users. BERT is incredible for its versatility and performance and because we used Nvidia's CUDA development tools alongside BERT, we were able to achieve mind blowing performance with just a single GPU (summarizing within seconds). We also used Google's AI Cloud services and customized them to the core. With the help of Google AI and BERT, we were able to detect how confident the doctor is in his diagnosis, something that computational AI has never achieved before. IntelliMed can also tell you the meaning of complicated words that may be difficult to understand using Google AI and our own Neural Networks. All of this was built in a mere 36-48 hours by a couple of high schoolers. To deliver such a great solution, we sure did run into a few problems. One of the more noteworthy ones would be deployment since neither of us were familiar with Google Cloud's deployment service and we couldn't use Linode since it required a credit card and we're still in high school :(. Deep Learning and summarization using BERT, confidence meter for doctor's diagnosis We learnt to work with Google Cloud services for the first time. Flask was also a new technology for us to learn but we found it to be easy since we were already familiar with Django (another web framework). IntelliMed has a plethora of new potential ideas for implementation. First, would-be full-blown accessibility features for speech, braille, and sound. We would achieve this by using Microsoft or Google's fantastic cloud accessibility features (that we already have buried in the code but are not using yet :) ). Second, would be to get multilingual support for multiple languages. And, last but not the least, integration with health apps on smartphones for deeper connectivity and an appointment system. This is what we envision IntelliMed's near future to be like.        
