I am currently a third-year Ph.D. student at the Department of Civil Engineering, Indian Institute of Science, Bangalore.
Python implementations and computational experiments accompanying the book "Mathematical Foundations of Deep Learning". This repository contains rigorous implementations of deep learning algorithms, mathematical concepts, optimization methods, neural network architectures, and theoretical demonstrations discussed throughout the book.
Sourangshu Ghosh Readme File
What if you could imitate a famous celebrity's voice or sing like a famous singer? This project started with a goal to convert someone's voice to a specific target voice. So called, it's voice style transfer. We worked on this project that aims to convert someone's voice to a famous English actress Kate Winslet's voice. We implemented a deep neural networks to achieve that and more than 2 hours of audio book sentences read by Kate Winslet are used as a dataset.
burgers_time_viscous, a FENICS code by Sourangshu Ghosh which uses the finite element method to solve a version of the time-dependent viscous Burgers equation over the interval [-1,+1].
Bitcoin price prediction algorithm using bayesian regression techniques
A rigorous, encyclopedic reference on the mathematical foundations of Deep Learning and Artificial Intelligence, organized as interlinked concept pages with precise definitions, formulations, and limitations.
Github Pages template for academic personal websites, forked from mmistakes/minimal-mistakes
A PIC/FLIP fluid simulation based on the methods found in Robert Bridson's "Fluid Simulation for Computer Graphics"
The Cuthill-Mckee algorithm is used for the reordering of a symmetric square matrix. It is based on the Breadth-First Search algorithm of a graph, whose adjacency matrix is the sparsified version of the input square matrix.The ordering is frequently used when a matrix is to be generated whose rows and columns are numbered according to the numbering of the nodes. By an appropriate renumbering of the nodes, it is often possible to produce a matrix with much smaller bandwidth. The Sparsified version of a matrix is a matrix in which most of the elements are zero.The Reverse Cuthill-Mckee Algorithm is the same as the Cuthill-Mckee algorithm, the only difference is that the final indices obtained using the Cuthill-Mckee algorithm are reversed in the Reverse Cuthill-Mckee Algorithm.
Research Paper: "Bayesian Beer Market Estimation: Simulating Nash Equilibrium Market Outcomes with Bayesian Analysis of Choice-Based Conjoint Data"
A Fortran Code able to generate Quantum Vortices which are topological effects(a group of spins that have a different topology than spins that point in only one direction) playing an important role in certain classical phase transitions
100 Computer Science Papers listed by Sourangshu Ghosh
Gaussian processes (GPs) are a good choice for function approximation as they are flexible, robust to over-fitting, and provide well-calibrated predictive uncertainty. Deep Gaussian processes (DGPs) are multi-layer generalisations of GPs, but inference in these models has proved challenging. Existing approaches to inference in DGP models assume approximate posteriors that force independence between the layers, and do not work well in practice. We present a doubly stochastic variational inference algorithm, which does not force independence between layers. With our method of inference we demonstrate that a DGP model can be used effectively on data ranging in size from hundreds to a billion points. We provide strong empirical evidence that our inference scheme for DGPs works well in practice in both classification and regression.
SEISMIC_CPML is a set of sixteen open-source Fortran90 programs to solve the two-dimensional or three-dimensional isotropic or anisotropic elastic, viscoelastic or poroelastic wave equation using a finite-difference method with Convolutional or Auxiliary Perfectly Matched Layer (C-PML or ADE-PML) conditions
SVGD is a general purpose variational inference algorithm that forms a natural counterpart of gradient descent for optimization. SVGD iteratively transports a set of particles to match with the target distribution, by applying a form of functional gradient descent that minimizes the KL divergence.
In numerical linear algebra, the Cuthill–McKee algorithm (CM), named for Elizabeth Cuthill and James McKee, is an algorithm to permute a sparse matrix that has a symmetric sparsity pattern into a band matrix form with a small bandwidth. The reverse Cuthill–McKee algorithm (RCM) due to Alan George is the same algorithm but with the resulting index numbers reversed
Annulus_flow, a FENICS code by Sourangshu Ghosh which models the flow of a fluid, governed by the time dependent Navier-Stokes equations, in a 2D eccentric annulus.
Computational mechanics for rough surfaces and fractures
A Simple Finite Element Method program
The objective of this work by Sourangshu Ghosh was to develop machine learning methods that could accurately predict adverse drug reactions using the SIDER and OFFSIDEs databases.
This is a collection of Fortran routines written by Sourangshu Ghosh over the years for use in more complex codes. The different files are as independent as possible from each other, but in some cases dependencies are unavoidable.
A Python software able to calculate the cross-section properties of combined steel sections
Simulating coronavirus with the SIR model
An autoencoder is a type of artificial neural network used to learn efficient data codings in an unsupervised manner.The aim of an autoencoder is to learn a representation (encoding) for a set of data, typically for dimensionality reduction, by training the network to ignore signal “noise”. Along with the reduction side, a reconstructing side is learnt, where the autoencoder tries to generate from the reduced encoding a representation as close as possible to its original input, hence its name. Several variants exist to the basic model, with the aim of forcing the learned representations of the input to assume useful properties. Examples are the regularized autoencoders (Sparse, Denoising and Contractive autoencoders), proven effective in learning representations for subsequent classification tasks,and Variational autoencoders, with their recent applications as generative models. Autoencoders are effectively used for solving many applied problems, from face recognition to acquiring the semantic meaning of words
Bayesian optimization is a method of finding the maximum of expensive cost functions. Bayesian optimization employs the Bayesian technique of setting a prior over the objective function and combining it with evidence to get a posterior function. This permits a utility-based selection of the next observation to make on the objective function, which must take into account both exploration (sampling from areas of high uncertainty) and exploitation (sampling areas likely to offer improvement over the current best observation).
A Python code useful for determining the Reliability of Weld Joints in Operational Nuclear Reactors
A C++ software useful for generating the seismic ground motion with inputs of Distance, fault dimensions, orientation and soil profile on origin and site
This is an implementation of IBM's Quantum Experience in simulation; a 5-qubit quantum computer with a limited set of gates.
Gaussian Process Motion Planner
allen_cahn_ode, a Python code by Sourangshu Ghosh which sets up and solves the 1D Allen-Cahn reaction-diffusion ordinary differential equation (ODE).