The task is to implement Machine Translator. An attention mechanism was used to deal with longer sequences. After data cleaning and processing, output labels were padded with start and end tokens before feeding to n/w.
Source CodeThe objective is to find the correct answer for the given question and context pair. Implemented Standford Attentive Reader. SQUAD v1 dataset was used for this project. Various binary and NLP features were used to get the best results. Compared the final results with fine-tuned BERT model.
Source CodeThe objective was for the given movie and the user to predict the rating given by him/her to the movie. The dataset was obtained from Kaggle. Matrix factorization was used to get similarity matrices. Tried and tested various ML models to get minimum Root Mean Square.
Source CodeThe objective is to predict as many as tags possible with high Precision and Recall. The dataset was obtained from Kaggle. The given problem is a multi-label classification problem. The dataset contains features such as Id, Title, Body and Tags. Data preprocessing and cleaning were done to remove HTML tags and hyperlinks. Micro-Averaged F1-Score was used as a performance metric as mentioned on Kaggle.
Source Code