Background: Growing evidence shows that fixed, nonpersonalized daily step goals can discourage individuals, resulting in unchanged or even reduced physical activity. Objective: The aim of this randomized controlled trial (RCT) was to evaluate the efficacy an automated mobile phone–based personalized and adaptive goal-setting intervention using machine learning as compared with active control steady 10,000. Methods: In 10-week RCT, 64 participants were recruited via email announcements...

As AI systems are integrated into high stakes social domains, researchers now examine how to design and operate them in a safe ethical manner. However, the criteria for identifying diagnosing safety risks complex contexts remain unclear contested. In this paper, we vagueness debates about behavior of systems. We show cannot be resolved through mathematical formalism alone, instead requiring deliberation politics development as well context deployment. Drawing from new sociotechnical lexicon,...

In many sequential decision-making settings where there is uncertainty about the reward of each action, frequent selection specific actions may reduce expected while choosing less frequently selected could lead to an increase. These effects are commonly observed in ranging from personalized healthcare interventions and targeted online advertising. To address this problem, authors propose a new class models called ROGUE (reducing or gaining unknown efficacy) multiarmed bandits. paper, present...

<sec> <title>BACKGROUND</title> Many mental health conditions (e.g., substance use or panic disorders) involve long-term patient assessment and treatment. Growing evidence suggests that the progression presentation of these may be highly individualized. Digital sensing predictive modeling can augment scarce clinician resources to expand personalize care. This manuscript discusses techniques process data into risk predictions, for instance lapse a with alcohol disorder (AUD). Of particular...

We here present some opinions about the challenge of understanding what newer forms artificial intelligence (AI) are, and how they are similar to or different from human (HI)in its many manifestations. Our central thesis is this: now that inner workings an AI may consist a huge network, billions weight parameters, we might really benefit research framework studies these new intelligences using array tools have been developed assess dimensions aspects in humans animals. feel current model...

Regular physical activity is associated with reduced risk of chronic illnesses. Despite various types successful interventions, maintenance over the long term extremely challenging.The aims this original paper are to 1) describe engagement post intervention, 2) identify motivational profiles using natural language processing (NLP) and clustering techniques in a sample women who completed 3) compare sociodemographic clinical data among these identified cluster groups.In cross-sectional...

The implementation of AI systems has led to new forms harm in various sensitive social domains. We analyze these as problems How address harms remains at the center controversial debate. In this paper, we discuss inherent normative uncertainty and political debates surrounding safety systems.of vagueness illustrate shortcomings current technical approaches Safety literature, crystallized three dilemmas that remain design, training deployment systems. argue resolving render a system 'safe'...

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

Designing systems with human agents is difficult because it often requires models that characterize agents' responses to changes in the system's states and inputs. An example of this scenario occurs when designing treatments for obesity. While weight loss interventions through increasing physical activity modifying diet have found success reducing individuals' weight, such programs are maintain over long periods time due lack patient adherence. A promising approach increase adherence...

Many multi-agent systems have the structure of a single coordinator providing behavioral or financial incentives to large number agents. Two challenges faced by are finite budget from which allocate incentives, and an initial lack knowledge about utility function Here, we present analytics approach solve coordinator's problem when agents make decisions maximizing functions that depend on prior system states, inputs, other parameters initially unknown subject temporal dynamics. Our framework...

As AI systems become prevalent in high stakes domains such as surveillance and healthcare, researchers now examine how to design implement them a safe manner. However, the potential harms caused by stakeholders complex social contexts address these remains unclear. In this paper, we explain inherent normative uncertainty debates about safety of systems. We then problem vagueness examining its place design, training, deployment stages system development. adopt Ruth Chang's theory intuitive...

Many settings involve sequential decision-making where a set of actions can be chosen at each time step, action provides stochastic reward, and the distribution for reward is initially unknown. However, frequent selection specific may reduce its expected while abstaining from choosing an cause to increase. Such non-stationary phenomena are observed in many real world such as personalized healthcare-adherence improving interventions targeted online advertising. Though finding optimal policy...

Distributed model predictive control (MPC) is either cooperative or competitive, and control-theoretic properties have been less studied in the competitive (e.g., game theory) setting. This paper studies MPC with linear dynamics a Stackelberg structure: Given fixed lower-level (LoMPC) controller, bilevel (BiMPC) controller chooses inputs to steer LoMPC knowing that optimizing respect different cost function. After defining BiMPC, we give examples demonstrate how interconnections dynamic can...

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 address this that training surrogate explainer which approximate how more complex model computing its predictions. Explainer are generally classified as either local or global explainers depending on what portion data space they purported explain. The improved coverage usually comes at expense...

Controlling heating, ventilation, and air conditioning (HVAC) to improve its energy efficiency or implement ancillary services (e.g., demand response frequency regu- lation) requires compensating for occupants the computers/equipment they use, which significantly impact thermal dynamics of buildings. Several studies have explored use either binary models online learning occupancy within model predictive control (MPC) frameworks, but more fine-grained has been less well-studied. This paper...

Human-machine systems, especially those involving reinforcement learning (RL), are becoming increasingly common across application domains. Designing these systems to be effective and reliable requires a task-oriented understanding of both human (HL) RL. In particular, how does the structure task affect performance humans RL algorithms? Games other environments can serve as important tools in this line research. While trend toward complex has led improved capabilities, such difficult use for...

Artificial Neural Networks (ANN) have been shown to be effective for many predictive tasks, such as system identification and reinforcement learning. However, they become more ubiquitous, there several examples of models exhibiting anthropomorphic bias (e.g. making predictions correlated with race or gender unrelated tasks) due over fitting, amplifying systematizing already inherent in training data. To address this problem, we consider a novel regularization approach deep learning, inspired...

Diabetes is a global health priority, especially in low- and-middle-income countries, where over 50% of premature deaths are attributed to high blood glucose. Several studies have demonstrated the feasibility using Community Health Worker (CHW) programs provide affordable and culturally tailored solutions for early detection management diabetes. Yet, scalable models design implement CHW while accounting screening, management, patient enrollment decisions not been proposed. We introduce an...

Multi-armed bandit models have proven to be useful in modeling many real world problems the areas of control and sequential decision making with partial information. However, scenarios, such as those prevalent healthcare operations management, maker's expected reward will decrease if an action is selected too frequently while it may recover they abstain from selecting this action. This scenario further complicated when choosing a particular also expends random amount limited resource where...