A Unifying Framework for Combining Complementary Strengths of Humans and ML toward Better Predictive Decision-Making

Abstract

Hybrid human-ML systems are increasingly in charge of consequential decisions in a wide range of domains. A growing body of work has advanced our understanding of these systems by providing empirical and theoretical analyses. However, existing empirical results are mixed, and theoretical proposals are often incompatible with each other. Our goal in this work is to bring much-needed organization to this field by offering a unifying framework for understanding conditions under which combining complementary strengths of human and ML leads to higher quality decisions than those produced by them individually-a state to which we refer to as human-ML complementarity. We focus specifically on the context of human-ML predictive decision-making systems and investigate optimal ways of combining human and ML-based predictive decisions, accounting for the underlying causes of variation in their judgments. Within this scope, we present two crucial contributions. First, drawing upon prior literature in human psychology, machine learning, and human-computer interaction, we introduce a taxonomy characterizing a wide variety of criteria across which human and machine decision-making differ. Building on our taxonomy, our second contribution presents a unifying optimizationbased framework for formalizing how human and ML predictive decisions should be aggregated optimally. We show that our proposed framework encompasses several existing models of human-ML complementarity as special cases. Last but not least, the exploratory analysis of our framework offers a critical piece of insight for future work in this area: the mechanism by which we combine human-ML judgments should be informed by the underlying causes of their diverging decisions. * indicates equal contribution. The ordering was decided by a coin toss.

Designing for Appropriate Reliance: The Roles of AI Uncertainty Presentation, Initial User Decision, and User Demographics in AI-Assisted Decision-Making

“…We speculate that this may be because humans with contextual knowledge (see Appendix C) are better at certain cases that may involve domain shifts. Prior works have shown that deep neural networks are prone to fail under dataset shifts [36,68,72], and lack the contextual knowledge and commonsense reasoning that humans have [50,73]. Our conjecture aligns with a prior exploration that observed amateur players-machine teaming can out-perform machines alone and grand-masters alone in chess [42].…”

A Note on Information Bias and Efficiency of Composite Likelihood

“…There have been efforts to model human-machine collaborations from both the fields of economics and computer science. Rastogi et al [2022] and Donahue et al [2022] investigate the optimal ways to aggregate independent algorithmic and human predictions, and under what conditions these aggregations outperform individual predictions. Straitouri et al [2022] proposes an algorithm that selects an optimal set of possible labels and presents it to a human expect for final selection.…”

Algorithmic Decision-Making Safeguarded by Human Knowledge

“…This research is broadly related to two streams of literature: those papers providing conceptual or theoretical frameworks for human-AI collaboration and empirical papers documenting real-world interactions between AI and human analysts. In the first stream, recent literature in computer science aims at the optimal integration of human and AI decisions (Bansal et al 2021, 2019, Donahue et al 2022, Gao et al 2021, Keswani et al 2021, Madras et al 2018, Mozannar and Sontag 2020, Raghu et al 2019, Rastogi et al 2022, Wilder et al 2020). On the one hand, Madras et al (2018) propose a learning-to-defer framework in which the AI can choose to make decision by its own or just pass the task to the downstream human expert.…”

“…Follow-up papers extend the framework to more complex settings, such as multiple experts (Keswani et al 2021), bandit feedback (Gao et al 2021), joint optimization of the prediction algorithm and pass function (Mozannar andSontag 2020, Wilder et al 2020). On the other hand, Donahue et al (2022), Rastogi et al (2022) consider a weighted average aggregation of human and AI decisions and show conditions for human-AI complementarity in which the aggregated decision outperforms both individual decisions. Recently, Grand-Clément and Pauphilet (2022) show that in the setting of sequential decision-making, the AI algorithm should be trained differently when a human analyst is involved.…”