A straightforward introduction to Vector Quantized Generative Adversarial Networks Some days ago, I was trying to understand a brand new architecture, MaskGit, a Masked Generative Image Transformer based on VQ-GAN. Thus, I searched a short summary on Medium about the topic, but, apparently, VQ-GAN were only discussed in relation to…

Presented with short summaries — Recently, I have started collaborating with the team of MarkTechPost, a California-based Artificial Intelligence news platform. My role is to choose two papers each month, which I consider fundamental, and write a short, easy-to-understand summary. I have decided to post these summaries (to be honest, a summary of these summaries)…

A brief resume of two years of my primary Ph.D. research — Table of contents Introduction Background The full pipeline of our system Results Conclusion References 1. Introduction Since the beginning of my Ph.D., I have been collaborating with the orthopedic team of the CTO (Center for Orthopaedic Trauma) of Turin (Italy), to develop an algorithm able to assist physicians in…

A user-friendly explanation of the brand new architecture introduced by Google — Table of contents Introduction The basis of CNNs From Self-Attention to Vision Transformers CoAtNet Conclusion References 1. Introduction On September 15, 2021, a new architecture defined the state-of-the-art (SOTA) in the ImageNet competition. CoAtNet (pronounced “coat” net) achieved a 90.88% top-1 accuracy with the huge JFT-3B dataset. More importantly, CoAtNet, when pre-trained with the (relatively)…