人类-人工智能团队何时能超越个体?具有不可能保证的紧界
When Can Human-AI Teams Outperform Individuals? Tight Bounds with Impossibility Guarantees
AI总结:
研究通过整合信号检测理论与信息论分析,推导出信心基于聚合规则的紧界,揭示互补性实现的条件及限制,提供设计公式并验证在多类任务中的有效性。
摘要:
人类-人工智能团队在70%的研究中未能超越其最佳成员,但没有任何理论能说明互补性何时可实现。我们通过整合信号检测理论与信息论分析,推导出以置信度为基础的聚合规则的紧界,得出四个结果:(1)互补性定理(团队超越个体当且仅当误差相关性ρ_HM < ρ*,其中ρ*≈a在对称近偶然 regime 中);(2)极小极大界显示收益随元认知敏感度差异按Θ(√Δd)缩放;(3)不可能结果证明当ρ_HM ≥ ρ*时,没有任何置信度为基础的聚合规则能实现互补性;(4)多类泛化ρ_K*≈ρ*/√(K-1)。预测匹配观察到的团队准确性(ImageNet-16H上R=0.94,CIFAR-10H上R=0.91)以及多类阈值缩放在人类数据上成立(R=0.93,K=16),在非高斯分布下具有鲁棒性。该框架解释了互补性为何罕见,并提供了可操作的设计公式;结果适用于聚合,不适用于生成新答案的交互性 deliberation。
英文摘要:
Human-AI teams fail to outperform their best member in 70% of studies, yet no theory specifies when complementarity is achievable. We derive tight bounds for the broad class of confidence-based aggregation rules by integrating signal detection theory with information-theoretic analysis, yielding four results: (1) a complementarity theorem (teams outperform individuals iff error correlation $ρ_{HM} < ρ^*$, with $ρ^* \approx a$ in the symmetric near-chance regime); (2) minimax bounds showing gains scale as $Θ(\sqrt{Δd})$ with metacognitive sensitivity difference; (3) an impossibility result proving no confidence-based aggregation rule achieves complementarity when $ρ_{HM} \geq ρ^*$; and (4) multi-class generalization $ρ^*_K \approx ρ^*/\sqrt{K-1}$. Predictions match observed team accuracy ($R = 0.94$ on ImageNet-16H, $R = 0.91$ on CIFAR-10H) and the multi-class threshold scaling holds on human data ($R = 0.93$, $K = 16$), with robustness under non-Gaussian distributions. The framework explains why complementarity is rare and provides actionable design formulas; results apply to aggregation, not to interactive deliberation that generates novel answers.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。