EffPPL is a shallowly embedded domain-specific probabilistic programming library in Multicore OCaml made using effect handlers.
Implements the Hamiltonian Monte Carlo Algorithm for performing Bayesian inference to obtain comparable performance against Stan.
Link to GitHub Repository can be found here.
Link to the detailed report and presentation can be found here.
Given preferences of elements, in an even sized set, it finds a stable matching of elements
Implemented a C++ program to find possible solutions of the stable roommate problem.
It uses Irving’s Algorithm which performs at O(n2) complexity.
Link to GitHub Repository can be found here
Worked on improving the work of Amin et. al.'s iPromoter-BnCNN for promoter classification.
Improved the accuracy of their constitutional architecture from 88% to 90%, using a variety of Deep Learning techniques.
Report along with the runnable source code can be found here.
Wrote a compiler for a subset of the C language using flex and bison.
It supports functions, conditional statements and iterative statements.
Wrote a compiler which does optimisations like constant propagation, common sub-expression elimination etc. for a smaller subset of the language.
The executable for the same can be found here.
Thoroughly read and understood the NeurIPS 2019 Paper.
Video Presentation along with Ravi Gupta can be found here.
Involves constructing an HACK micro-architecture using a Hardware Simulator
Implemented a HACK assembler to convert mnemonics to Machine Language.
Implemented a virtual machine and a compiler, which converts JACK (High-level Language) to mnemonics.
Link to GitHub Repository can be found here
Got a detailed understanding of the Linux's Energy Aware Scheduler .
Proposed some algorithmic improvements to the existing Energy Aware Scheduler.
A detailed report of the work done can be found here.
Wrote a mini-interpreter for Prolog using OCaml.
The interpreter is capable of supporting backtracking and choice points.