TUPLE-Bench: Evaluating and Mitigating Privacy-Utility Trade-offs in Typed Decision Models

Paper recorded by Signals 4 on 2026-10-16 in cs.AI. Abstract reproduced from arXiv; link to the original below.

Published 2026-10-16 on arXiv · recorded by Signals 4 on 2026-10-19

Category: cs.AI · 人工智能 · first seen 2026-10-19

Abstract

Typed decision models often return more than a selected category. They also expose probabilities for each candidate and, sometimes, a separate confidence score. These fields support downstream decisions but may reveal information beyond a recipient's authorization. We introduce TUPLE-Bench (Typed decision models' Utility and Privacy Leakage Evaluation Benchmark), a benchmark of paired synthetic ca

Read on arXiv →