Home
The problem of unlabelled data is a big issue in machine learning. With an increase in the availability of clean data we are able to train more accurate models for better performance and broaden the variety of machine learning problems attempted. There is a lot of data on the internet that is not labelled and it is a tedious job which most people do not like to do. In this project, we have designed an online website that allows clients to upload a set of images that they would like to have labelled. In addition to uploading the images, the client is required to attach a “fee” they are willing to pay upon completion of the task. A potential labeller is able to select which tasks they want to work on based on the set of images provided and the specified remuneration.