Reproduce “MithraDetective: A System for Cherry-picked Trendlines Detection” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yoko Nagafuchi, Yin Lin, Kaushal Mamgain, Abolfazl Asudeh, H. V. Jagadish, Cong Yu
MithraDetective, developed by researchers from the University of Michigan, University of Illinois at Chicago, and Google Research, introduces a systematic approach to detect cherry-picked trendlines by formally defining and calculating the 'support' for data-driven claims. The system provides an algorithmic foundation and user-friendly interface to evaluate the statistical validity of trend interpretations, demonstrating its utility with real-world datasets like COVID-19 and climate change.
Key insights
- Subjective trend interpretation can be transformed into objective statistical evaluation through the concept of 'support,' quantifying how well a statement is represented by the data.
- A formal framework for trendline statements and their 'support' calculation allows for systematic detection of cherry-picked data.
- Multiple algorithmic approaches (exact, optimized exact, random sampling) can efficiently calculate trend support across diverse dataset sizes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.