Reinforcement Learning Under Algorithmic Triage

Straitouri, Eleni; Singla, Adish; Meresht, Vahid Balazadeh; Gomez-Rodriguez, Manuel

doi:10.48550/arxiv.2109.11328

Cited by 3 publications

(4 citation statements)

References 22 publications

(25 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For our approach, in addition to cases of sensing failure, we consider the task of delegation [ 26 , 27 , 28 , 29 , 30 , 31 , 32 ]. We are primarily focused on the manager’s ability to take advantage of higher-performing agents in a given context.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Compensating for Sensing Failures via Delegation in Human–AI Hybrid Systems

Fuchs

Passarella

Conti

2023

Sensors

View full text Add to dashboard Cite

Given the increasing prevalence of intelligent systems capable of autonomous actions or augmenting human activities, it is important to consider scenarios in which the human, autonomous system, or both can exhibit failures as a result of one of several contributing factors (e.g., perception). Failures for either humans or autonomous agents can lead to simply a reduced performance level, or a failure can lead to something as severe as injury or death. For our topic, we consider the hybrid human–AI teaming case where a managing agent is tasked with identifying when to perform a delegated assignment and whether the human or autonomous system should gain control. In this context, the manager will estimate its best action based on the likelihood of either (human, autonomous) agent’s failure as a result of their sensing capabilities and possible deficiencies. We model how the environmental context can contribute to, or exacerbate, these sensing deficiencies. These contexts provide cases where the manager must learn to identify agents with capabilities that are suitable for decision-making. As such, we demonstrate how a reinforcement learning manager can correct the context–delegation association and assist the hybrid team of agents in outperforming the behavior of any agent working in isolation.

show abstract

Section: Related Workmentioning

confidence: 99%

“…For instance, ref. [ 27 ] demonstrates a method that partially inspired our approach. In their work, they define and train a reinforcement learning model that can operate under algorithmic triage.…”

Section: Related Workmentioning

confidence: 99%

Compensating for Sensing Failures via Delegation in Human–AI Hybrid Systems

Fuchs

Passarella

Conti

2023

Sensors

View full text Add to dashboard Cite

show abstract

“…Cognitive scientists have developed and tested different theories about the cognitive process underpinning responsibility judgments (Alicke, 2000;Chockler & Halpern, 2004;Gerstenberg & Lagnado, 2010;Shaver, 2012). However, the increasing development of AI systems that assist and collaborate with humans, rather than replacing them (Balazadeh Meresht et al, 2022;De et al, 2020De et al, , 2021Mozannar et al, 2022;Okati et al, 2021;Raghu et al, 2019;Straitouri et al, 2021;Wilder et al, 2021), calls for more empirical and theoretical research to shed light on the way humans make responsibility judgments in situations involving human-AI teams (Cañas, 2022). Recent work in that area has identified several factors that influence responsibility judgments (Awad et al, 2020;Lima et al, 2021;Longin et al, 2023).…”

Section: Introductionmentioning

confidence: 99%

Towards a computational model of responsibility judgments in sequential human-AI collaboration

Tsirtsis,

Rodriguez,

Gerstenberg

2024

Preprint

View full text Add to dashboard Cite

When a human and an AI agent collaborate to complete a task and something goes wrong, who is responsible? Prior work has developed theories to describe how people assign responsibility to individuals in teams. However, there has been little work studying the cognitive processes that underlie responsibility judgments in human-AI collaborations, especially for tasks comprising a sequence of interdependent actions. In this work, we take a step towards filling this gap. Using semi-autonomous driving as a paradigm, we develop an environment that simulates stylized cases of human-AI collaboration using a generative model of agent behavior. We propose a model of responsibility that considers how unexpected an agent’s action was, and what would have happened had they acted differently. We test the model’s predictions empirically and find that in addition to action expectations and counterfactual considerations, participants’ responsibility judgments are also affected by how much each agent actually contributed to the outcome.

show abstract

“…Learning under algorithmic triage seeks the development of machine learning models that operate under different automation levels-models that take decisions for a given fraction of instances and leave the remaining ones to human experts [8,24,25,26,27]. This line of work has predominantly focused on supervised learning settings with a few very recent notable exceptions [28,29]. However, in this line of work, each sample is either predicted by the model or by the human expert.…”

Section: Introductionmentioning

confidence: 99%

Provably Improving Expert Predictions with Prediction Sets

Straitouri¹,

Wang²,

Okati³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Automated decision support systems promise to help human experts solve tasks more efficiently and accurately. However, existing systems typically require experts to understand when to cede agency to the system or when to exercise their own agency. Moreover, if the experts develop a misplaced trust in the system, their performance may worsen. In this work, we lift the above requirement and develop automated decision support systems that, by design, do not require experts to understand when to trust them to provably improve their performance. To this end, we focus on multiclass classification tasks and consider automated decision support systems that, for each data sample, use a classifier to recommend a subset of labels to a human expert. We first show that, by looking at the design of such systems from the perspective of conformal prediction, we can ensure that the probability that the recommended subset of labels contains the true label matches almost exactly a target probability value. Then, we identify the set of target probability values under which the human expert is provably better off predicting a label among those in the recommended subset and develop an efficient practical method to find a near-optimal target probability value. Experiments on synthetic and real data demonstrate that our system can help the experts make more accurate predictions and is robust to the accuracy of the classifier it relies on.

show abstract

Reinforcement Learning Under Algorithmic Triage

Cited by 3 publications

References 22 publications

Compensating for Sensing Failures via Delegation in Human–AI Hybrid Systems

Compensating for Sensing Failures via Delegation in Human–AI Hybrid Systems

Towards a computational model of responsibility judgments in sequential human-AI collaboration

Provably Improving Expert Predictions with Prediction Sets

Contact Info

Product

Resources

About