Beyond the Gold Standard in Analytic Automated Essay Scoring
Gabrielle Gaudeau · 2025
Automated Essay Scoring (AES) has shifted from holistic to analytic approaches, enabling more trait-based feedback generation. However, analytic scoring is cognitively demanding and bias-prone. Current systems collapse rater disagreements into single labels, obscuring genuine variability. We propose a novel system that learns from individual raters, improving trustworthiness and formative feedback quality.