One of the best papers I came across in 2020 was the Google Brain team’s Self-training with Noisy Student. It is a semi-supervised technique that works well when we have an insufficient amount of labeled data. So much so that it beat the contemporary... [Read More]

I have given a loose introduction to active learning in my previous post.. In this post, I’ll dig deeper into the application of active learning in computer vision. To understand the problem better, imagine you are asked to prepare a model for classification on... [Read More]

Supervised learning is great. You provide data and labels and the model will be trained. The state of the art in computer vision is almost solely based on supervised learning. But there’s a cost that’s associated with labeling of data that usually simplified in terms of time and monetary resources... [Read More]

The Sortobot project is aimed at the development of indigenous technology for sorting and grading of agro-commodities at an industrial scale. It uses computer vision to analyze the surface of instances of the commodities and uses machine learning for classification based on various user-selectable parameters such as the severity of... [Read More]

The Agrograde project brings the power of AI in the hands of marginalized farmers and small traders who are particularity vulnerable to consignment rejection and the resulting financial losses due to disputes over quality parameters of the vegetables and fruits. This app enables farmers to snap an image of randomly... [Read More]