Showcase of different assignment projects done through my undergraduate study.

Higgs Boson Detection Simplified

Motivation

Identify the Higgs boson from sanitized data provided by CERN using a binary classifier

Description

• Build a lightweight framework to adapt the number of layers and nodes of a neural network

Results

Obtained the highest score

Convergence based on the training/test dataset (Top). Classifier results on the unknown dataset with a score of 82.5% accuracy (Bottom).

Origin of a Nuclear Accident

Motivation

Identify the origin of a nuclear accident from gas propagation through a simplified setting.

Description

• Implement a simple fluid simulation
• Build an optimization scheme

Results

A simple gas simulation.

3 different data were collected through 3 sensors based on the best parameters found through my optimization (right). A visualization of every optimization step through each sub-region (left).

nu-Bootis Orbital Parametrization

Motivation

Improve convergence of a genetic algorithm to find the best parameters of the Kepler formula to fit observed data.

Description

• Implement a custom strategy for a genetic algorithm

Results

Classical genetic algorithm to find the minima of a 2D sinc function.

Optimization score across each iteration, comparing different optimization strategies (Left). Convergence of optimization, consistent with the Kepler formula, based on artificially sampled points. (Right).

Dark Matters

Motivation

Enhance galaxy simulation by incorporating a dark matter halo to match experimental observation of gravitational potential.

Description

• Implement a small N-Body simulation
• Estimate the gravitational potential of a simulated galaxy

Results

Dark matter contribution to galaxy rotation. Each blue dot represents a star. Each line in red is an isocontour of the gravitational potential.