Welcome to My Portfolio
ABOUT MYSELF
I am Asswin C R, currently pursuing a Master's in Artificial Intelligence at Nanyang Technological University (NTU), Singapore. I hold a Bachelor's degree from Amrita Viswa Vidyapeetham, Coimbatore, in Computer Science with a specialization in Artificial Intelligence, where I graduated with a First Class and Distinction.
My passion lies in Deep Machine Learning, Computer Vision, Natural Language Processing and Data Analytics . I am dedicated to leveraging these interests to contribute effectively to the field of AI.
RELEVANT COURSEWORK
Artificial Intelligence - Deep Learning, Python for Machine Learning, Natural Language Processing, Computer Vision, Deep Reinforcement Learning, Big Data Analysis, Speech & Image Processing, Robotics Operating System, Robotic Vision.
Programming - Data Structures and Algorithms, Design Analysis and Algorithms.
Mathematics - Linear Algebra, Probability, Probabilistic Graphical Models, Statistics, Markov and Conditional Random Fields (MRF & CRF), Financial Time Series Analysis.
TECHNICAL SKILLS
Programming Languages - Python, C++, Java, MATLAB, R, HTML/CSS, SQL.
Libraries - TensorFlow, Keras, PyTorch, Scikit Learn, OpenCV, OpenAI gym, Spacy.
Tools and Frameworks - Linux, Git, Android Studio, Apache Spark, Jupyter Notebook, Anaconda.
Project & Experience
Pediatric Pneumonia Diagnosis using Stacked Ensemble Learning on Architectures on Multi-Model Deep CNN Architectures
• The work proposes Contrast Limited Adaptive Histogram Equalization(CLAHE) for image enhancement and a stacking classifier based on the fusion of deep learning-based features for pediatric pneumonia diagnosis.
• The stacking classifier validated using Stratified K-Fold cross-validation achieves an accuracy of 98.62%, precision of 98.99%, recall of 99.53%, F1-score of 99.26%, and an AUC score of 93.17% on the publicly available pediatric pneumonia dataset.
Clinical Text Classification (Natural Language Processing)
• Performed Sequence & Non-Sequence Modeling for classifying clinical/medical transcriptions prescribed by the doctor.
• In sequence modeling, LSTM and Transformers models use word embedding and positional word embedding respectively. In non-sequence modeling, ML models such as Logistic regression, multi-class classifier with KNN, and random forest were used.
Big Data Analysis on Crime Dataset
Analysis on various aspects of Crime such as Narcotics, Assault, Burglary etc and plotting them on World Map.(Tools Used: Scala , Apache Spark , Python , Microsoft Excel).
Fully Connected Tensor Network
• Applying HOSVD on Micro seismic data and Image Reconstruction using FTCN Algorithm.
• Tool used : MATLAB and Python
Citation Context Classification
• Classified the citations from different datasets, using various ML algorithms and text pre-processing techniques.
• The result analysis concludes Multinomial naive Bayes Algorithm works best for this project(context classification) followed by KNN and Random Forest. The least-performing algorithms are Decision Tree and Support Vector Machines.
Markov Random Fields for Image Segmentation & Denoising
• For Image segmentation, we find the pair-wise and unary potential function and maximize each pixel individually.
• For Image denoising, we find the Clique potential and unary potential and maximize its probabilty.
Applying Dense Neural Network for Vehicles classification
Implemented DNN for Vehicles Image Classification and did various analysis on the architecture to get better accuracy(91.52%).
Internship
Machine Learning Internship-Technocolabs
Stock Market Prediction
• Developed a Machine Learning algorithm to predict if the price of a stock market incurs a profit or loss, given its current O/H/C/L/V values for a given date and deploying it on the Heroku server.
• Tools or techniques used: Python, Django, Flask, Heroku
Robotic Automation Intern -Yaskawa India Private Limited
Stock Market Prediction
• Programmed and simulated arc welding and palletizing robots using MotoSim EG-VRC & pendant.
• Modeled and analyzed 3D robotics palletizing algorithm and mechanics for mixed carton sizes.
• Developed a C++ algorithm to effectively use robot’s memory while data transfer.