Equipment downtime is a significant challenge for many industries. In oil extraction, downtime costs can be as high as $250 000 per day. To prevent downtime, technicians manually interact with the equipment or monitor its health using sensory signals. Sensory data indirectly ascertain equipment health, while manual actions (inspections or repairs) provide a direct and precise insight but are time-consuming and costly. Thus, efficiently leveraging sensory data and outcomes of manual actions to accurately estimate the health of their equipment while finding the critical time points to schedule repairs and minimize the overall downtime is a crucial challenge faced by industries. In this article, we present a novel joint modeling approach called the action specific-input output hidden Markov model (AS-IOHMM) that integrates real-time sensor data and discrete health state information obtained by manual actions to aid prognosis and decision making of industrial equipment. In contrast to existing models that assume nondecreasing degradation without considering maintenance actions, AS-IOHMM estimates the impact of different maintenance actions on equipment health by learning action-specific transition probability matrices. We assess the effectiveness of AS-IOHMM through a numerical case study and validate its performance using mud-pump maintenance and sensory data from an oil rig, demonstrating enhanced prognosis ability and cost reduction of 7–15% over existing methods.

A key challenge for decision makers when incorporating black box machine learned models into practice is being able to understand the predictions provided by these models. One set of methods proposed to address this challenge is that of training surrogate explainer models which approximate how the more complex model is computing its predictions. Explainer methods are generally classified as either local or global explainers depending on what portion of the data space they are purported to explain. The improved coverage of global explainers usually comes at the expense of explainer fidelity (i.e., how well the explainer's predictions match that of the black box model). One way of trading off the advantages of both approaches is to aggregate several local explainers into a single explainer model with improved coverage. However, the problem of aggregating these local explainers is computationally challenging, and existing methods only use heuristics to form these aggregations. In this paper, we propose a local explainer aggregation method which selects local explainers using non-convex optimization. In contrast to other heuristic methods, we use an integer optimization framework to combine local explainers into a near-global aggregate explainer. Our framework allows a decision-maker to directly tradeoff coverage and fidelity of the resulting aggregation through the parameters of the optimization problem. We also propose a novel local explainer algorithm based on information filtering. We evaluate our algorithmic framework on two healthcare datasets: the Parkinson's Progression Marker Initiative (PPMI) data set and a geriatric mobility dataset from the UCI machine learning repository. Our choice of these healthcare-related datasets is motivated by the anticipated need for explainable precision medicine. We find that our method outperforms existing local explainer aggregation methods in terms of both fidelity and coverage of classification. It also improves on fidelity over existing global explainer methods, particularly in multi-class settings, where state-of-the-art methods achieve 70% and ours achieves 90%.

The implementation of AI systems has led to new forms of harm in various sensitive social domains. We analyze these as problems How to address these harms remains at the center of controversial debate. In this paper, we discuss the inherent normative uncertainty and political debates surrounding the safety of AI systems.of vagueness to illustrate the shortcomings of current technical approaches in the AI Safety literature, crystallized in three dilemmas that remain in the design, training and deployment of AI systems. We argue that resolving normative uncertainty to render a system 'safe' requires a sociotechnical orientation that combines quantitative and qualitative methods and that assigns design and decision power across affected stakeholders to navigate these dilemmas through distinct channels for dissent. We propose a set of sociotechnical commitments and related virtues to set a bar for declaring an AI system 'human-compatible', implicating broader interdisciplinary design approaches.

Despite recent interest in both the critical and machine learning literature on "bias" in artificial intelligence (AI) systems, the nature of specific biases stemming from the interaction of machines, humans, and data remains ambiguous. Influenced by Gendler's work on human cognitive biases, we introduce the concept of alief-discordant belief, the tension between the intuitive moral dispositions of designers and the explicit representations generated by algorithms. Our discussion of alief-discordant belief diagnoses the ethical concerns that arise when designing AI systems atop human biases. We furthermore codify the relationship between data, algorithms, and engineers as components of this cognitive discordance, comprising a novel epistemic framework for ethics in AI.