{"ID":2832203,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.06341","arxiv_id":"2512.06341","title":"Interpretive Efficiency: Information-Geometric Foundations of Data Usefulness","abstract":"Interpretability is central to trustworthy machine learning, yet existing metrics rarely quantify how effectively data support an interpretive representation. We propose Interpretive Efficiency, a normalized, task-aware functional that measures the fraction of task-relevant information transmitted through an interpretive channel. The definition is grounded in five axioms ensuring boundedness, Blackwell-style monotonicity, data-processing stability, admissible invariance, and asymptotic consistency. We relate the functional to mutual information and derive a local Fisher-geometric expansion, then establish asymptotic and finite-sample estimation guarantees using standard empirical-process tools. Experiments on controlled image and signal tasks demonstrate that the measure recovers theoretical orderings, exposes representational redundancy masked by accuracy, and correlates with robustness, making it a practical, theory-backed diagnostic for representation design.","short_abstract":"Interpretability is central to trustworthy machine learning, yet existing metrics rarely quantify how effectively data support an interpretive representation. We propose Interpretive Efficiency, a normalized, task-aware functional that measures the fraction of task-relevant information transmitted through an interpreti...","url_abs":"https://arxiv.org/abs/2512.06341","url_pdf":"https://arxiv.org/pdf/2512.06341v1","authors":"[\"Ronald Katende\"]","published":"2025-12-06T08:11:22Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.IR\",\"cs.IT\"]","methods":"[]","has_code":false}