Crowdsourcing: Classification Methodology

This lecture covers the concept of crowdsourcing, where tasks are distributed to a large group of people. It explains the process of submitting tasks, collecting answers, and different types of crowd-workers. The instructor discusses non-iterative and iterative aggregation algorithms, such as Majority Decision and Expectation Maximisation, used to aggregate answers from workers. The lecture also delves into the Honey Pot method, which identifies unreliable workers. Finally, it explores the Expectation Maximisation algorithm in detail, emphasizing its steps in estimating labels and worker expertise.