Fact-Checking at Scale with DimensionRank
This addresses the challenge of misinformation for users of major internet platforms, offering a novel approach that is incremental in improving upon existing one-dimensional systems.
The paper tackles the problem of fact-checking at scale on internet platforms like web search and social media by proposing a platform-based solution with a two-dimensional rating system (agreement and hotness), which enables new information-seeking queries and supports formal proofs in propositional calculus.
The most important problem that has emerged after twenty years of popular internet usage is that of fact-checking at scale. This problem is experienced acutely in both of the major internet application platform types, web search and social media. We offer a working definition of what a "platform" is. We critically deconstruct what we call the "PolitiFact" model of fact checking, and show it to be inherently inferior for fact-checking at scale to a platform-b ased solution. Our central contribution is to show how to effectively platformize the problem of fact-checking at scale. We show how a two-dimensional rating system, with dimensions agreement and hotness allows us to create information-seeking queries not possible with the on e-dimensional rating system predominating on existing platforms. And, we show that, underlying our user-friendly user-interface, lies a system that allows the creation of formal proofs in the propositional calculus. Our algorithm is implemented in our open-source DimensionRank software package available at "https://thinkdifferentagain.art".