Research-led explanations of human capacity, learning and technological change. Start with the foundations, then follow the evidence and related ideas as the Learning Hub grows.
Bottom-up attention describes attentional priority driven strongly by properties of incoming stimuli rather than by an observer’s current goal. The concept remains useful, but modern attention research shows that real selection rarely divides cleanly into bottom-up and top-down causes.
The Brier score evaluates probabilistic forecasts by comparing stated probabilities with outcomes. It is a strictly proper scoring rule and can support confidence-calibration research, but a single Brier score combines calibration with discrimination or resolution, so it should not be mistaken for a pure measure of whether confidence matches accuracy.