Stella

Abstract

Geotagged data (e.g. images or news items) have empowered various important applications, e.g., search engines and news agencies. However, the lack of available geotagged data significantly reduces the impact of such applications. Meanwhile, existing geotagging approaches rely on the existence of prior knowledge, e.g., accurate training dataset for machine learning techniques. This paper presents Stella; a crowdsourcing framework for image geotagging. The high accuracy of Stella is resulted by being able to recruit workers near the image location even without knowing its location. In addition, Stella also return its confidence about the reported location to help users in understanding the result quality. Experimental evaluation shows that Stella consistently geotags an image with an average of 95% accuracy and 90% of confidence.

Comparing Generic and Community-Situated Crowdsourcing for Data Validation in the Context of Recovery from Substance Use Disorders

“…While online crowdsourcing markets make it convenient to pay for workers willing to solve a range of diferent tasks, they suffer from limitations such as not attracting enough workers with desired background or skills [4,21,54]. For example, it can be a challenge to recruit workers for a task that requires workers who speak a specifc language or who live in a certain city [10,42] . Situated crowdsourcing can help fll in the gaps in these scenarios where the crowd needs to be associated with some context.…”

“…It allows for a geo-fenced and more contextually controlled crowdsourcing environment, thus enabling targeting certain individuals, leveraging people's local knowledge or cognitive states, or simply reaching an untapped source of potential workers [27][28][29]34]. Researchers have discussed benefts of targeting geographically or temporally situated crowds over generic crowds in scenarios like providing emergency services in disasters [56], geotagging photos [42], etc. An experiment by Ipeirotis et al automatically identifed "situated crowds" with desired competence to complete a task and demonstrated that the cost of hiring workers through their platform is less than that of hiring workers through paid crowdsourcing platforms [38].…”

“…Attracting workers who are likely to have the skills needed for the target task may be useful to overcome this problem. Previous research has shown efectiveness of crowdsourcing with the "right workers" (e.g., who are present at a particular location or time, or have familiarity with or expertise in performing particular tasks) [5,17,42,86]. Building and expanding on this idea, we aim to understand if the unique perspectives and contextual knowledge from members of a community relevant to the target task can help achieve better quality results in crowdsourcing.…”