PinnedShashank KumarinTowards Data ScienceAI for painting: Unraveling Neural Style TransferThe post discusses the mathematics behind neural style transfer, an algorithm that helps make artistic changes to images.Feb 5, 2022Feb 5, 2022
PinnedShashank KumarinTowards Data ScienceImage Segmentation, UNet, and Deep Supervision Loss Using Keras ModelDeep CNNs used for segmentation often suffer from vanishing gradients. Can we combat this by calculating loss at different output levels?Sep 28, 2022Sep 28, 2022
Shashank KumarinTowards Data ScienceWhy WGANs beat GANs: A journey from KL divergence to Wasserstein lossWasserstein distance helps WGANs outperform vanilla GANs and VAEs. This post explains why so using some easy math.Jan 12, 2023Jan 12, 2023
Shashank KumarinTowards Data ScienceHow to de-blur images without training neural networksAre you bad at photography and click blurry pictures? This post is a guide on how to use Fourier transform to unblur images.Nov 29, 2022Nov 29, 2022
Shashank KumarinTowards AIEvaluating Mode Collapse in GANs Using NDB ScoreGANs frequently slip into the fathomless pit that is mode collapse. Let’s learn to quantitatively track it using NDB score.Nov 17, 2022Nov 17, 2022
Shashank KumarinTowards Data ScienceWhat can Kevin Durant teach about the probability of a probability?Let’s use Kevin Durant’s three-point stat line from last Sunday’s game to understand beta distributions.Jun 17, 2021Jun 17, 2021
Shashank KumarinTowards Data ScienceChi-squared tests to compare two machine learning models and determine whether they are randomMay 18, 20211May 18, 20211
Shashank KumarinCantor’s ParadiseFermat, Descartes, and The Dawn of Differential CalculusBackgroundMay 17, 2021May 17, 2021
Shashank KumarinTowards Data ScienceA very Bayesian interpretation of decision trees and other machine learning algorithmsMay 13, 2021May 13, 2021
Shashank KumarinTowards Data ScienceBuilding the optimization problem of maximum margin classifiersBackgroundApr 2, 2021Apr 2, 2021