Projects

My Work

Here are some of the projects I have worked on:

• Transfer Learning:
  
  • Demonstrated Transfer Learning by fine-tuning a pre-trained ResNet18 model on CIFAR-10 dataset.
  • Model was trained by freezing it completely and then training its outer layers to compare accuracies.



• Classification with BERT:
  
  • Developed a fake news classification model by fine-tuning a BERT model.
  • Utilized the BERT base model with ~110 million parameters and trained models through PEFT learning, achieving remarkable accuracies compared to traditional models.



• Semi-Supervised Learning:
  
  • Employed an unsupervised clustering algorithm on a vast dataset, utilizing labels from a small subset
  • With an optimal cluster estimation, achieved a rapid model with commendable accuracy.



• Spam Classification:
  
  • Developed a machine learning model achieving high accuracy in distinguishing between spam and non-spam messages.
  • Utilized various established ML algorithms & conducted an analysis of their performance metrics for comparison.



• EDA on Flight Dataset Using R:
  
  • Conducted in-depth exploratory data analysis (EDA) on flight data using R, ggplot2, and R Markdown to unveil patterns in flight delays, identify prominent routes, and detect outliers.
  • Deployed an interactive app with RShiny to visualize & present the findings for enhanced accessibility & engagement.