AI ResearchEnvironmental Research
May 20, 2026AI Research

Epistemic Humility in Neural Network Interpretability

Current interpretability methods reveal correlations between model internals and human-legible concepts, but conflating correlation with causal explanation leads to overconfident claims about how these systems actually reason.

Machine LearningPhilosophy of ScienceAI
February 14, 2026Environmental Research

Microclimate Variation in Dense Urban Blocks

Street-level temperature measurements across 47 city blocks reveal that microclimate variation within a single neighborhood frequently exceeds the variation between neighborhoods — with significant implications for heat-risk mapping.

ClimateUrban StudiesEnvironmental Science