Projects
Fake News Detection
- Analysed the LIAR dataset in a team of 2, consisting of 12.8K manually labelled short statements from various contexts
- Engineered algorithms such as CNN, LSTM, SVM based on surface-level linguistic patterns (Python)
- Accomplished accuracy of 60% for binary classification and 27% for six way multiclass classification
Compiler Design and Construction
- Developed a toy compiler for ERPLAG language specifications (C Programming)
- Implemented modules include lexical Analyzer, Syntax Analyzer, Semantic Analyzer, and Code generator
Cross Language Information Retrieval
- Implemented IBM models (1, 2) for developing a translator using English and Dutch corpora (Python)
- Improved convergence rate of EM algorithm to find the translational probabilities, achieving 85.4% accuracy
Part-Of-Speech Tagger
- Implemented Viterbi algorithm in hidden Markov model, learning from tagged dataset to tag unknown data (Python)
- Achieved accuracy of 88% and 86% correspondingly on English and Chinese datasets
Traffic Officer Scooters Application
- Maximized activity points by placing officers in non-conflicting locations, satisfying constraints on grid coordinates
- Engineered algorithms such as Depth First Search, A* Search, Backtracking to optimize placement of officers (Python)
Search Engine
- Implemented Spell Check, AutoComplete and Snippet features in the search queries (Apache Solr, PHP)
IC Tester in Assembly
- Designed a microprocessor based tester to test the logical functioning of Integrated Chips(IC)
- Developed an assembly language program and displayed result as PASS or FAIL on LCD
- Tested multiple IC - 7408, 7486, 7432, simulated using Proteus software
RAM using Dynamic Sets
- Developed a simple abstraction of a computer's RAM using multiple dynamic sets along with defragmentation
- Separated the filled and empty nodes in the defragmentation process
- Implemented using array data structures in C programming