- As Andrew Ng says now "AI is the new electricity", I guess I am totally powered by it now.
- I spent my Summer'19 at Niramai (Banglore) where I worked on Breast Cancer Detection using Deep Learning which led to a research paper under review at San Antonio Breast Cancer Symposium(SABCS)'19.
- Me and my team attended I-Hack-2019, IIT Bombay (Jan'19) where we presented our project on automated Abusive/Bad sentiment comment flagging on Youtube using Deep Learning techniques and we won first prize(INR 1Lakh).
- Implemented research papers on Generative Adversarial Networks (GANs) for fake face(CelebA dataset) and fake digit generation(MNIST dataset)
- Since my freshman year, I am undertaking projects which uses ML/DL to solve real life problems. From designing basic regression problems to training Deep Convolutional Neural Nets on platforms like PyTorch and Keras, its been a great journey till now. By the end of my first year, I participated in DotSlash Hackathon(NIT-Surat) where my team implemented a speech recognition system using Azure API and further participated in its second edition too with another project combining DL and fuzzy-logic techniques.
- Lately I am getting more inclined towards Computer Vision and NLP. I undertook one project where I extracted tweets from the twitter for the sentiment analysis, and ultimately used 1-D Convolutional layers coupled with Embedding layers (on both character and word level) to classify the sentiment of the tweet. This project made me read a lot of research papers which is certainly a plus for me.
Projects
Sentinel : An AI that watches your back
The project focuses on automating the deletion of unwanted comments during a live stream which hampers user expeience while Digital Streaming.jQuery, AJAX, Flask, Keras, Android Studio, Deep Learning, REST APIStudent Intervention System (S.I.S)
Our project focuses on improving the Student Evaluation Critirea by intervening at the right time by evaluating his socio economic conditions and also appreciate if they perform relatively well.Flask, Keras, Java, REST API, MATLABSkills
Python
PyTorch
Keras
NLP
Computer Vision
Experience
- Niramai - Machine Learning InternMay 2019 - July 2019
*** Paper under review at San Antonio Breast Cancer Symposium(SABCS)’19 ***
- Breast Density plays a major role in Mammographic assessment for detecting breast cancer. Highly dense breasts hinders a lot of information in mammographic images making it ultimately useless for drawing conclusions.
- Designed the entire pipeline for estimation of breast density from breast thermal images.
- Explored various state-of-the-art deep learning models as baseline for further improvements
- Enhanced the overall specificity/sensitivity ratio by obtaining results from ensemble of many techniques.
- Tested multi cascade architecture with shared weights for multi view-multi input analysis of breast density