Suboptimal Behavior Research Articles

Interference effects in commercially available free-space silicon single-photon avalanche diodes

Single-photon avalanche diodes (SPADs) are essential for photon-based measurements and metrology, enabling measurement comparisons at the few-photon level and facilitating global traceability to the SI. A spatially uniform detector response is crucial for these applications. Here, we report on interference effects in commercially available silicon SPADs that are detrimental to their spatial uniformity. Contrasts as high as 18% are observed, posing problems for metrology and general applications that utilize coherent light and require stable detection efficiencies. We eliminate the device optical window as a contributing interface, isolating likely causes to anti-reflective coatings, the semiconductor surface, and the SPAD's internal structure. We also present results where we leverage this sub-optimal behavior by aligning an incident beam with the position of maximum constructive interference, yielding an effective detection efficiency of 51.1(1.7)% compared to the normal value of 44.3(1)% obtained with the interference suppressed. We anticipate that this work will significantly impact the continuing development of these devices, the methods for characterizing them, and their use in accurate measurements.

Formal contracts mitigate social dilemmas in multi-agent reinforcement learning

Multi-agent Reinforcement Learning (MARL) is a powerful tool for training autonomous agents acting independently in a common environment. However, it can lead to sub-optimal behavior when individual incentives and group incentives diverge. Humans are remarkably capable at solving these social dilemmas. It is an open problem in MARL to replicate such cooperative behaviors in selfish agents. In this work, we draw upon the idea of formal contracting from economics to overcome diverging incentives between agents in MARL. We propose an augmentation to a Markov game where agents voluntarily agree to binding transfers of reward, under pre-specified conditions. Our contributions are theoretical and empirical. First, we show that this augmentation makes all subgame-perfect equilibria of all Fully Observable Markov Games exhibit socially optimal behavior, given a sufficiently rich space of contracts. Next, we show that for general contract spaces, and even under partial observability, richer contract spaces lead to higher welfare. Hence, contract space design solves an exploration-exploitation tradeoff, sidestepping incentive issues. We complement our theoretical analysis with experiments. Issues of exploration in the contracting augmentation are mitigated using a training methodology inspired by multi-objective reinforcement learning: Multi-Objective Contract Augmentation Learning. We test our methodology in static, single-move games, as well as dynamic domains that simulate traffic, pollution management, and common pool resource management.

Model-driven design and validation of glucose supply control in a tubular photobioreactor operated under oxygen-balanced mixotrophy

Scaling up bioprocesses remains challenging due to the partially unpredictable and suboptimal behavior of microbes during scale transition. Mathematical modelling serves as a valuable tool in navigating these challenges. In this study, we developed and validated a kinetic model based on the tank-in-series approach for a novel bioprocess known as oxygen-balanced mixotrophy with the microalga Galdieria sulphuraria. This model aims to facilitate the design of an effective glucose feeding control strategy tailored for two-phase tubular photobioreactors (TPBRs). The residence time distribution of our reference reactor was investigated by means of a tracer experiment. Qualitative analysis of the results indicated that our system was optimally represented by simulations with a 100 stirred-tank reactors (STRs).The controlled variable, oxygen concentration in the gas phase, showed a downward trend in simulations with simplified conditions: fixed oxygen production and balanced glucose supply. The decrease was caused by dominant net heterotrophic activity and accumulation of CO2 in the gas phase. Different configurations of a proportional-integral-derivative (PID) controller were designed by means of the ultimate gain method and compared under these simplified conditions. The most effective PID controller was implemented and refined in simulations with outdoor weather conditions. The model was validated successfully by simulating over time in a lab-scale STR the spatial fluctuations happening in a TPBR. The empirical oxygen consumption and production were faster and steeper than in the model, probably due to inaccuracies in parameter estimation or biomass adaptation.

Alternatives to Bayesian Updating

We discuss models of updating that depart from Bayes’ rule even when it is well-defined. After reviewing Bayes’ rule and its foundations, we begin our analysis with models of non-Bayesian behavior arising from a bias, a pull toward suboptimal behavior due to a heuristic or a mistake. Next, we explore deviations caused by individuals questioning the prior probabilities they initially used. We then consider non-Bayesian behavior resulting from the optimal response to constraints on perception, cognition, or memory, as well as models based on motivated beliefs or distance minimization. Finally, we briefly discuss models of updating after zero probability events.

Investigating an effort avoidance account of attentional strategy choice.

People often choose suboptimal attentional control strategies during visual search. This has been at least partially attributed to the avoidance of the cognitive effort associated with the optimal strategy, but aspects of the task triggering such avoidance remain unclear. Here, we attempted to measure effort avoidance of an isolated task component to assess whether this component might drive suboptimal behavior. We adopted a modified version of the Adaptive Choice Visual Search (ACVS), a task designed to measure people's visual search strategies. To perform optimally, participants must make a numerosity judgment-estimating and comparing two color sets-before they can advantageously search through the less numerous of the two. If participants skip the numerosity judgment step, they can still perform accurately, albeit substantially more slowly. To study whether effort associated with performing the optional numerosity judgment could be an obstacle to optimal performance, we created a variant of the demand selection task to quantify the avoidance of numerosity judgment effort. Results revealed a robust avoidance of the numerosity judgment, offering a potential explanation for why individuals choose suboptimal strategies in the ACVS task. Nevertheless, we did not find a significant relationship between individual numerosity judgment avoidance and ACVS optimality, and we discussed potential reasons for this lack of an observed relationship. Altogether, our results showed that the effort avoidance for specific subcomponents of a visual search task can be probed and linked to overall strategy choices.

Don't Give-Up: Why some intervention schemes encourage suboptimal behavior.

Many social challenges stem from individuals' tendency to prefer immediately rewarding but suboptimal behaviors ("Give-Up" options) over more costly endeavors that yield much better outcomes in the long run ("Try" options). For example, many people forgo the long-term benefits of formal education, healthy diets, learning new technologies, and even finding true love. This paper examines various incentivization programs that combine external rewards and punishments to discourage such counterproductive behaviors, which often result in only temporary behavioral change. Our findings suggest that some interventions' limited impact may be due to their focus on only shifting behaviors from "Give-Up" (e.g., dropping out of college, avoiding the gym) to "Try" (e.g., attending college, exercising regularly), without promoting sufficient exploration of these "Try" options. Yet exploration of the long-term benefits of "Trying" may be crucial to increase the chances of long-term learning and commitment. Using a simplified abstraction of this dilemma, our results show a high tendency to choose "Give-Up" options prior to intervention. Examination of four different incentivization strategies suggests that onlyrewarding exploration of new "Try" options is a straightforward strategy to increase exploration and optimal choice. Punishing both the selection of "Give-Up" options and the choice to exploit suboptimal "Try" options produced similar results. Other common guidance strategies were less effective, as these strategies simply tended to replace one suboptimal behavior with another. Surprisingly, punishments seemed to bea relatively more successful incentive than rewards. We discuss how these insights can help guide policy aiming to improve long-term outcomes through incentivization.

Caregiver greeting to infants under 6 months already reflects emerging differences in those later diagnosed with autism.

As infants develop, caregivers adjust their behaviour to scaffold their infant's emerging skills, such that changes in infants' social abilities are expected to elicit changes in caregiver behaviour. We examined whether changes in the probability of infant-directed caregiving behaviour-specifically, greeting, a ubiquitous signal used by caregivers to initiate reciprocal interactions-differ between infant-caregiver dyads with an infant later diagnosed with autism and dyads with a neurotypically developing infant during infants' first 6 months. Using longitudinal data from 163 dyads, we found that caregivers in autism dyads (n = 40) used greeting less and at later infant ages than caregivers with a neurotypically developing infant (neurotypical dyads, n = 83). Caregivers in dyads with infants at elevated familial genetic likelihood for autism who did not receive an autism diagnosis (EL-non-autism dyads, n = 40) showed no differences in greeting compared with neurotypical dyads. Socioeconomic status partially mediated the difference between autism and neurotypical dyads. These findings show that autism and socioeconomic status were associated with the mutually adapted dynamics of dyadic interaction beginning in the first postnatal weeks. Importantly, differences in caregiver greeting observed in autism dyads are not interpreted as suboptimal behaviour from caregivers but rather indicate how early emerging social differences related to autism, years before overt features are present, may alter social learning opportunities elicited by the infant.

Alternative translation initiation by ribosomal leaky scanning produces multiple isoforms of the Pif1 helicase.

In budding yeast, the integrity of both the nuclear and mitochondrial genomes relies on dual-targeted isoforms of the conserved Pif1 helicase, generated by alternative translation initiation (ATI) of PIF1 mRNA from two consecutive AUG codons flanking a mitochondrial targeting signal. Here, we demonstrate that ribosomal leaky scanning is the specific ATI mechanism that produces not only these, but also novel, previously uncharacterized Pif1 isoforms. Both in-frame, downstream AUGs as well as near-cognate start codons contribute to the generation of these alternative isoforms. This has crucial implications for the rational design of genuine separation-of-function alleles and provides an explanation for the suboptimal behaviour of the widely employed mitochondrial- (pif1-m1) and nuclear-deficient (pif1-m2) alleles, with mutations in the first or second AUG codon, respectively. We have taken advantage of this refined model to develop improved versions of these alleles, which will serve as valuable tools to elucidate novel functions of this helicase and to disambiguate previously described genetic interactions of PIF1 in the context of nuclear and mitochondrial genome stability.

Self-care of patients with coronary heart disease: a commentary on various suboptimal behaviours highlighted by a longitudinal study.

Self-care of patients with coronary heart disease: a commentary on various suboptimal behaviours highlighted by a longitudinal study.

An Approach to the Systematic Characterization of Multitask Accelerated CNN Inference in Edge MPSoCs

Deep Learning is ubiquitous today and is increasingly moving from the cloud down to the edge of networked infrastructures, where it enables embedded applications to perform complex inference tasks close to the data sources, reducing long-distance data movement and alleviating the need for a powerful cloud infrastructure. Edge-class multi-processor system on chip (MPSoC) devices featuring an on-chip FPGA fabric offer key advantages for Deep Learning inference tasks, especially for complex applications where multiple models may be run concurrently in the same platform. In this work, we propose an approach and a practical framework for the systematic characterization of multithreaded Deep Learning inference on edge FPGA MPSoCs. We instantiate the framework into a real-world MPSoC platform, targeting Xilinx Vitis-AI as a representative example of a commercial Deep Learning acceleration toolkit for edge environments. We design a comprehensive experimental campaign and apply it to the platform for several convolutional neural networks, each trained on three different datasets. We show that our approach can be used for both hardware- and software-level analysis of a target system. Among other findings, the analysis revealed a suboptimal behavior of the underlying toolkit runtime, involving the utilization of the accelerator cores and the uneven software latency of the support library, influenced by the shapes of the input tensors.

Analysis of patient safety event report categories at one large academic hospital.

Given the persistent safety incidents in operating rooms (ORs) nationwide (approx. 4,000 preventable harmful surgical errors per year), there is a need to better analyze and understand reported patient safety events. This study describes the results of applying the Team Strategies and Tools to Enhance Performance and Patient Safety (TeamSTEPPS) supported by the Teamwork Evaluation of Non-Technical Skills (TENTS) instrument to analyze patient safety event reports at one large academic medical center. Results suggest that suboptimal behaviors stemming from poor communication, lack of situation monitoring, and inappropriate task prioritization and execution were implicated in most reported events. Our proposed methodology offers an effective way of programmatically sorting and prioritizing patient safety improvement efforts.

Potential-Based Reward Shaping for Intrinsic Motivation (Student Abstract)

Recently there has been a proliferation of intrinsic motivation (IM) reward shaping methods to learn in complex and sparse-reward environments. These methods can often inadvertently change the set of optimal policies in an environment, leading to suboptimal behavior. Previous work on mitigating the risks of reward shaping, particularly through potential-based reward shaping (PBRS), has not been applicable to many IM methods, as they are often complex, trainable functions themselves, and therefore dependent on a wider set of variables than the traditional reward functions that PBRS was developed for. We present an extension to PBRS that we show preserves the set of optimal policies under a more general set of functions than has been previously demonstrated. We also present Potential-Based Intrinsic Motivation (PBIM), a method for converting IM rewards into a potential-based form that are useable without altering the set of optimal policies. Testing in the MiniGrid DoorKey environment, we demonstrate that PBIM successfully prevents the agent from converging to a suboptimal policy and can speed up training.

Stitching Sub-trajectories with Conditional Diffusion Model for Goal-Conditioned Offline RL

Offline Goal-Conditioned Reinforcement Learning (Offline GCRL) is an important problem in RL that focuses on acquiring diverse goal-oriented skills solely from pre-collected behavior datasets. In this setting, the reward feedback is typically absent except when the goal is achieved, which makes it difficult to learn policies especially from a finite dataset of suboptimal behaviors. In addition, realistic scenarios involve long-horizon planning, which necessitates the extraction of useful skills within sub-trajectories. Recently, the conditional diffusion model has been shown to be a promising approach to generate high-quality long-horizon plans for RL. However, their practicality for the goal-conditioned setting is still limited due to a number of technical assumptions made by the methods. In this paper, we propose SSD (Sub-trajectory Stitching with Diffusion), a model-based offline GCRL method that leverages the conditional diffusion model to address these limitations. In summary, we use the diffusion model that generates future plans conditioned on the target goal and value, with the target value estimated from the goal-relabeled offline dataset. We report state-of-the-art performance in the standard benchmark set of GCRL tasks, and demonstrate the capability to successfully stitch the segments of suboptimal trajectories in the offline data to generate high-quality plans.

Pre-acquired Functional Connectivity Predicts Choice Inconsistency.

Economic choice theories usually assume that humans maximize utility in their choices. However, studies have shown that humans make inconsistent choices, leading to suboptimal behavior, even without context-dependent manipulations. Previous studies showed that activation in value and motor networks are associated with inconsistent choices at the moment of choice. Here, we investigated if the neural predispositions, measured before a choice task, can predict choice inconsistency in a later risky choice task. Using functional connectivity (FC) measures from resting-state functional magnetic resonance imaging (rsfMRI), derived before any choice was made, we aimed to predict subjects' inconsistency levels in a later-performed choice task. We hypothesized that rsfMRI FC measures extracted from value and motor brain areas would predict inconsistency. Forty subjects (21 females) completed a rsfMRI scan before performing a risky choice task. We compared models that were trained on FC that included only hypothesized value and motor regions with models trained on whole-brain FC. We found that both model types significantly predicted inconsistency levels. Moreover, even the whole-brain models relied mostly on FC between value and motor areas. For external validation, we used a neural network pretrained on FC matrices of 37,000 subjects and fine-tuned it on our data and again showed significant predictions. Together, this shows that the tendency for choice inconsistency is predicted by predispositions of the nervous system and that synchrony between the motor and value networks plays a crucial role in this tendency.

Mental Models and Learning: The Case of Base-Rate Neglect

We experimentally document persistence of suboptimal behavior despite ample opportunities to learn from feedback in a canonical updating problem where people suffer from base-rate neglect. Our results provide insights on the mechanisms hindering learning from feedback. Importantly, our results suggest mistakes are more likely to be persistent when they are driven by incorrect mental models that miss or misrepresent important aspects of the environment. Such models induce confidence in initial answers, limiting engagement with and learning from feedback. We substantiate these insights in an alternative scenario where individuals involved in a voting problem overlook the importance of being pivotal. (JEL D83, D91)

Preemptive patenting, secondary patents, and the persistence of monopoly

AbstractWe develop a theoretical economic model that establishes the link between fragmented patenting systems in an industry already characterized as a monopoly resulting in continuous/steady monopolistic market structure. We demonstrated that preemptive patenting strategy, by using secondary patenting scheme, may lead to an increase in the number of patents without changing actual number of products on the market. Moreover, when such a strategy is employed by a monopoly, it only serves to sustain and further monopolistic and therefore socially suboptimal market structure. Institutional/legal framework must exist to promote this kind of socially suboptimal behavior and resulting outcomes.

Adaptive trajectory-constrained exploration strategy for deep reinforcement learning

Deep reinforcement learning (DRL) faces significant challenges in addressing hard-exploration tasks with sparse or deceptive rewards and large state spaces. These challenges severely limit the practical application of DRL. Most previous exploration methods relied on complex architectures to estimate state novelty or introduced sensitive hyperparameters, resulting in instability. To mitigate these issues, we propose an efficient adaptive trajectory-constrained exploration strategy for DRL. The proposed method guides the agent’s policy away from suboptimal solutions by regarding previous offline demonstrations as references. Specifically, this approach gradually expands the exploration scope of the agent and strives for optimality in a constrained optimization manner. Additionally, we introduce a novel policy-gradient-based optimization algorithm that utilizes adaptive clipped trajectory-distance rewards for both single- and multi-agent reinforcement learning. We provide a theoretical analysis of our method, including a deduction of the worst-case approximation error bounds, highlighting the validity of our approach for enhancing exploration. To evaluate the effectiveness of the proposed method, we conducted experiments on two large 2D grid world mazes and several MuJoCo tasks. The extensive experimental results demonstrated the significant advantages of our method in achieving temporally extended exploration and avoiding myopic and suboptimal behaviors in both single- and multi-agent settings. Notably, the specific metrics and quantifiable results further support these findings. The code used in the study is available at https://github.com/buaawgj/TACE.

Optimal Timeout Choices in Clutch Situations in the NBA

When down by a possession in fourth quarter with the shot clock off and possessing the ball, National Basketball Association (NBA) coaches face a decision: take a timeout and draw up a play or let play continue. Utilizing play-by-play data for 16,839 NBA games, we find taking a timeout lowers a team's chances of a successful possession by 5.7 percentage points. Coaches might view taking a timeout as the safe decision, since blame for the loss would fall on the players for not executing the play, not on the coach for failing to call one. This suboptimal behavior suggests there is a principal–agent problem between team owners and coaches.

Assessing Predictors of Self-Care Behavior and Practices to Mitigate Air Pollution: A Cross-Sectional Study Among Undergraduate Students in Rawalpindi, Pakistan

Air pollution constitutes a pressing worldwide public health issue, and Pakistan is the fourth most polluted country, where nearly 99% of its populace breathes in the air quality that exceeds the WHO criteria. Embracing WHO-endorsed strategies can help in extending life expectancy by five years. This investigation was designed to delve into the factors influencing self-care behaviors and practices regarding air pollution based on Health Belief Model among undergraduate students. Methods: A questionnaire based on the Health Belief Model was used to conduct a cross sectional study on undergraduate students from Rawalpindi, Pakistan about their self-care and protection practices against air pollution. The questionnaire had seven sections and was validated. Non-probability convenience sampling was used. SPSS 26, Spearman’s correlation, and linear regression were used for data analysis. Results: The study enrolled 292 university students, with 189 (64.7%) being male and 103 (35.3%) females, and a mean age of 21.07±1.896. Findings revealed that 22.6% of respondents exhibited satisfactory self-care behavior, while 77.4% exhibited suboptimal behavior. In terms of self-care preventive practices, 58.6% scored poorly, with only 41.4% demonstrating good practices. Conclusions: The study's outcomes demonstrated poor self-care and protection practices among the participants. Factors such as perceived severity, perceived benefits, self-efficacy, and cues to action significantly influenced preventive behaviors. Whereas, perceived susceptibility and barriers were found insignificant. These findings underscore a notable association between Health Belief Model constructs and students' adoption of preventive measures against air pollution.

A suggestion-based hierarchical energy efficient control framework for connected and automated vehicles in mixed stream urban road networks

Connected and automated vehicles (CAVs) and their advanced control systems have the ability to improve energy efficiency and mobility of the transportation system, but hundred percent market penetration of CAVs will not happen in the near future. CAVs are thus expected to share the road with human-driven vehicles (HDVs), which can act in suboptimal manners. Since vehicles on road constitute an interactive system, where actions of one vehicle influence the actions of its surrounding vehicles, suboptimal behavior of HDVs affect all their surrounding vehicles including the CAVs. This paper thus focuses on the development of a suggestion-based control framework for CAVs in interactive mixed traffic environments. Our mixed traffic environment includes both CAVs and connected HDVs that have vehicle-to-vehicle (V2V) communication capabilities. In this suggestion-based control framework, a CAV is considered to provide velocity and lane-change suggestions to multiple surrounding HDVs to follow. These suggestions are the velocities and lane-change actions the CAV wants the HDVs to follow to improve its own and the group’s energy efficiency and mobility. To take into account the interactive nature of the environment, the CAV is considered to jointly plan its own trajectory, evaluate the suggestions to the surrounding HDVs, and predict the actions of its surrounding vehicles. The joint planning requires solving the problem in joint state- and action-space, and in this paper, we develop a Monte Carlo Tree Search (MCTS)-based trajectory planning approach for the CAV. Since the joint action- and state-space grows exponentially with the number of agents and can be computationally expensive, we propose an adaptive action-space by pruning the actions resulting in unsafe trajectories for each agent. The trajectory planning approach is followed by a low-level model predictive control (MPC)-based motion controller, which aims at tracking the reference trajectory in an optimal fashion. Simulation studies show the proposed control strategy’s efficacy compared with existing baseline methods. Simulation studies show improvements in fuel efficiency ranging from 4% to 50% for HDVs and from 1% to 30% for the CAV compared to the baseline.