A New Multi-task and Meta Reinforcement Learning Benchmark

Improvements in compute have been a key component in AI progress, thanks to Moore’s law or rather its generalization of continued exponential falling cost per unit of computation. Computational scaling can lead to better performance which is often complementary to the algorithm advances. To seek an improvement that makes a difference in the short term, AI researchers tend to utilize the human understanding of the domain, but it is often seen that deep learning methods rely less on human knowledge to produce dramatically more performance. Based on historical observations, building human-based knowledge into the agents didn’t help in the long…

A comprehensive analysis of Super-resolution Convolutional Neural Networks to benchmark Single Image Super-resolution

Super-resolution is a process of upscaling an image, by converting a given low-resolution image to a corresponding high-resolution image with better visual quality. High-resolution images provide improved reconstructed details of the scenes and constituent objects, or as they may state in any modern crime scene, enhance!

Super-resolution (SR) has applications in various domains including object detection in scenes, facial detection in surveillance videos, astronomical images, forensics, images in remote sensing, and medical imaging. It has been extensively studied due to its challenging nature.

Why is it a challenging problem?

Firstly, SR is an ill-posed inverse problem. This is an open research problem since it has multiple…

Ronak Mehta

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store