GUESS (Gaming Umbrella for Enterprise Social Sourcing) is a crowdsourcing and human computing game for the enterprise. GUESS leverages the playfulness of a game to harness "the wisdom of the crowd" in order to solve problems that computers still struggle with like image processing, speech recognition, and more. By completing simple tasks such as labeling images, transcribing speech, or annotating text, humans can help computers become smarter.
We created and deployed two instances of the GUESS game:
- GUESS what? – Building datasets and taxonomies
- Large enterprises often need to generate datasets for research, maintenance, and optimization purposes. Gathering the relevant information is both costly and time‐consuming.
- GUESS provides a mechanism that encourages masses of users to build datasets and hierarchical taxonomies. The datasets are built dynamically by large numbers of people and can be used to collect general or domain‐specific and community-biased information. The resulting datasets contain both the data and a ranking of its popularity, as reflected by the crowd's activity.
- GUESS who? – Filling gaps in social networks
- Social Networks & Discovery (SaND) is an aggregation platform for information discovery and analysis that leverages the complex relationships between content and people. It does this by mining data sources like bookmarks, tags, and intranet content. But to generate a real picture of the people in an enterprise, SaND needs information that is not written anywhere.
- GUESS helps to fill in the gaps in social networks by collecting information on people and their relationships to other people and entities. An exciting scoring function encourages validation of terms.
- GUESS uses SaND technology to define the group of people a user will 'play' on, and then feeds the knowledge collected (tags, relationships) back to SaND—generating more precise and complete knowledge about the people in an enterprise.