Jordan Boyd-Graber, Massimiliano Ciaramita, Markus Leippold, Thomas Diggelmann, Jannis Bulian
We have not lifted any functions out of this paper's repositories yet, so there is nothing we have run. If it links a repository, it is listed below.
We introduce climate-fever, a new publicly available dataset for verification of climate change-related claims. By providing a dataset for the research community, we aim to facilitate and encourage work on improving algorithms for retrieving evidential support for climate-specific claims, addressing the underlying language understanding challenges, and ultimately help alleviate the impact of misinformation on climate change. We adapt the methodology of fever [1], the largest dataset of artificially designed claims, to reallife claims collected from the Internet. While during this process, we could rely on the expertise of renowned climate scientists, it turned out to be no easy task. We discuss the surprising, subtle complexity of modeling real-world climate-related claims within the fever framework, which we believe provides a valuable challenge for general natural language understanding. We hope that our work will mark the beginning of a new exciting long-term joint effort by the climate science and ai community. Recently, new literature on algorithmic fact-checking has emerged, using machine learning and natural language understanding (nlu) to work on this problem from different angles. One influential framework that combines several of these aspects is fever [1]. It consists of a well-vetted dataset of human-generated claims and evidence retrieved from Wikipedia and a shared-task for evaluating implementations of claim validators. Given that the fever claims are artificially constructed, they may not share the characteristics of real-world
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2012.00614")
get_code_for_paper("2012.00614")
have("2012.00614")
Connect an agent — have() is free.