Share Profile
SYED AMMAR AHMED
Core Skills
- C, C++, C#, R
- JavaScript
- Python, Stata
- LaTex
- Maple
- MS Office
Experience
Internship / Volunteer Work
- Bazaar Technologies - Software Engineer
Publications / Creative Projects
- Visual Landmarks Recognition of Urban Structures using Convolutional Neural Network: Research Publication in IEEE conference. The publication happened in the conference ICoMET 2020 and is available on IEEE Xplore. The model used is already deployed on Heroku.
- Deep CNN to estimate Photometric Redshish and explore Feature Space through Visualization
- Research publication: Deblending Galaxy images using Generative Adversarial Network
Final Year Project
Project Title
Estimating Photometric Redshifts with Deep Convolutional Neural Networks
Description
We have developed a new architecture which we call "PhotoNet" to estimate photometric redshifts of galaxies on large-scale surveys. The photometric part means that data comes from photometric surveys which consist of images from a telescope through different filters. We have worked with five channels ugriz data coming from SDSS (Sloan Digital Sky Survey) where ugriz means bands capture ultraviolet, green, red, infrared and brightness. We have also visualized the output from intermediate layers of our network to understand its behavior and working. The future of modern cosmology is studying large-scale structures and this model is an important contribution in astronomy that relies on future large-scale photometric surveys.