Summer Internship 2022
Hi folks, so this summer I had one of the most wonderful summer internships one could desire of, thanks to bennett university I got a chance to Go to the National Univerity of Singapore (NUS) in Singapore and do my internship/training on the topic of Data Analytics using Deep Learning in collaboration with Hewlett Packard Enterprise (HPE)
During this internship, we learned from the best teachers the world has to offer in this field which were: Dr. TAN Wee Kek, Dr. LEK Hsiang Hui, Dr. Amirhassan Monajemi from NUS, and from HPE we had Mr. Shantanu Pandey. Their teaching methods were fantastic, and the way they explained each concept in utter detail was simply MARVELOUS.
the pedagogy was in such a way, that the day we were taught the theory of a subject the very next day we would be tossed around with the real-life problems to be solved using the theory that we learned in earlier classes. Between all of these, we end up boosting our knowledge by learning: the foundation of Data Analytics, Exploratory Data Analytics, Descriptive Statistical Measures, Python Data Science Libraries, Regression Analysis, Classification, Clustering, Association, Text Mining, Exploring and applying ANN, CNN and RNN, GD algorithms and training techniques of ANN, Prototyping and Pivoting to develop innovative real-world solutions using Deep Learning & Analytics with a business focus. That sounds a lot right, well it surely was quite a hectic month filled with knowledge & code. Thanks to our mentors for making it fun to learn.
During this period we were made to create a project which was applicable in real-world scenarios. So in a team of 5, We decided to create our project “Your shoulder” in this we built A live chatbot that takes input from the user. Three working models that take inputs. The motive is to determine whether the person is suffering from depression or a prevalent mental disorder. Our assistance will further help the patient connect to a verified psychologist.
Here is our workflow diagram:
For pre-processing we used Rescaling, GRAY Scale Conversion, Face Detection, Data Cleaning, Vectorization, Bow encoding, tokenization, padding, and embedding to process different kinds of data.
in last we achieved an accuracy of 0.64 in CNN emotion detection, 0.90 in logistic model mood detection & 0.67 in mental health detection
using these models we created a bot that we deployed on a local host all by ourselves without taking the help of any third-party software. Here's a glimpse of our project.
on the backend level, a report was being made which recorded the user's emotions, and things they said, and the machine predicted the output of the user's diagnosis. heres the GitHub link for the project https://github.com/Yoshibansal/Your-Shoulder-App
The whole internship surely helped in Strengthening & develops Deep Learning, Analytical, Critical Thinking, Innovation, and Collaboration skills.
Apart from the internship I even had fun exploring Singapore
-Garvit Bandil