arXiv subjects

Hsuan-Wei Lee

Publications and source records attributed to Hsuan-Wei Lee.

At least 19 recordsLinked to original sources

arXiv2026

Professional networks and the diffusion of clinical guidelines in opioid prescribing

Large and persistent differences in opioid prescribing across physicians and regions cannot be explained by patient characteristics or physician attributes alone. We developed a behavioral framework in which prescribing evolves through persistence, exposure to peers in professional networks, and heterogeneous responses to a common policy signal that varies with network centrality. Using nationwide Medicare Part D data from 2013 to 2020, covering more than two million physician-year observations, we tested three hypotheses implied by this framework. Physicians exposed to higher peer prescribing subsequently prescribe more; more central physicians reduce prescribing more following the introduction of the 2016 CDC guideline, with no evidence of differential pre-trends; and changes in peer prescribing are closely associated with changes in individual prescribing in the post-guideline period. By 2020, physicians at the 90th percentile of network centrality exhibited prescribing reductions 0.30 percentage points larger than those at the 10th percentile, with the gap widening steadily after the introduction of the CDC guideline. Together, these results indicate that opioid prescribing operates through professional networks, in which policy effects spread through connections and appear to be shaped by network position. This suggests that engaging highly connected physicians may help extend the reach of opioid stewardship programs. It also raises questions about how the burden and benefits of such targeting would be distributed across physicians and patients.

physics.soc-ph↗

arXiv2026

Dissipative Quantum Dynamics in Static Network with Different Topologies

We investigate the dissipative dynamics of quantum population and coherence among different network topologies of a quantum network using a quantum spin model coupled to a thermal bosonic reservoir. Our study proceeds in two parts. First, we analyze a small network of Ising spins embedded in a large dissipative bath, modeled via the Lindblad master equation, where temperature arises naturally from system-bath coupling. This approach reveals how network topology shapes quantum dissipative dynamics, providing a basis for controlling quantum coherence through tailored network structures. Second, we propose a mean-field approach that extends the network to larger scales and captures dissipative dynamics in large-scale networks, connecting network topology to quantum coherence in complex systems and revealing the sensitivity of quantum coherence to network structure. Our results highlight how dissipative quantum dynamics depend on network topology, providing insight into the coherent dynamics of entangled states in networks. These results may be extended to dynamics in complex systems such as opinion propagation in social models, epidemiology, and various condensed-phase and biological systems.

quant-ph↗

arXiv2026

How Exploration Breaks Cooperation in Shared-Policy Multi-Agent Reinforcement Learning

Multi-agent reinforcement learning in dynamic social dilemmas commonly relies on parameter sharing to enable scalability. We show that in shared-policy Deep Q-Network learning, standard exploration can induce a robust and systematic collapse of cooperation even in environments where fully cooperative equilibria are stable and payoff dominant. Through controlled experiments, we demonstrate that shared DQN converges to stable but persistently low-cooperation regimes. This collapse is not caused by reward misalignment, noise, or insufficient training, but by a representational failure arising from partial observability combined with parameter coupling across heterogeneous agent states. Exploration-driven updates bias the shared representation toward locally dominant defection responses, which then propagate across agents and suppress cooperative learning. We confirm that the failure persists across network sizes, exploration schedules, and payoff structures, and disappears when parameter sharing is removed or when agents maintain independent representations. These results identify a fundamental failure mode of shared-policy MARL and establish structural conditions under which scalable learning architectures can systematically undermine cooperation. Our findings provide concrete guidance for the design of multi-agent learning systems in social and economic environments where collective behavior is critical.

cs.MA↗

arXiv2025

Q-Learning-Driven Adaptive Rewiring for Cooperative Control in Heterogeneous Networks

Cooperation emergence in multi-agent systems represents a fundamental statistical physics problem where microscopic learning rules drive macroscopic collective behavior transitions. We propose a Q-learning-based variant of adaptive rewiring that builds on mechanisms studied in the literature. This method combines temporal difference learning with network restructuring so that agents can optimize strategies and social connections based on interaction histories. Through neighbor-specific Q-learning, agents develop sophisticated partnership management strategies that enable cooperator cluster formation, creating spatial separation between cooperative and defective regions. Using power-law networks that reflect real-world heterogeneous connectivity patterns, we evaluate emergent behaviors under varying rewiring constraint levels, revealing distinct cooperation patterns across parameter space rather than sharp thermodynamic transitions. Our systematic analysis identifies three behavioral regimes: a permissive regime (low constraints) enabling rapid cooperative cluster formation, an intermediate regime with sensitive dependence on dilemma strength, and a patient regime (high constraints) where strategic accumulation gradually optimizes network structure. Simulation results show that while moderate constraints create transition-like zones that suppress cooperation, fully adaptive rewiring enhances cooperation levels through systematic exploration of favorable network configurations. Quantitative analysis reveals that increased rewiring frequency drives large-scale cluster formation with power-law size distributions. Our results establish a new paradigm for understanding intelligence-driven cooperation pattern formation in complex adaptive systems, revealing how machine learning serves as an alternative driving force for spontaneous organization in multi-agent networks.

physics.soc-ph↗

arXiv2025

Context-sensitive norm enforcement reduces sanctioning costs in spatial public goods games

Uniform punishment policies can sustain cooperation in social dilemmas but impose severe costs on enforcers, creating a second-order free-rider problem that undermines the very mechanism designed to prevent exploitation. We show that the remedy is not a harsher stick but a smarter one. In a four-strategy spatial public-goods game we pit conventional punishers, who levy a fixed fine, against norm-responsive punishers that double both fine and cost only when at least half of their current group already cooperates. Extensive large scale Monte Carlo simulations on lattices demonstrate that context-sensitive punishment achieves complete defector elimination at fine levels 15\% lower than uniform enforcement, despite identical marginal costs per sanctioning event. The efficiency gain emerges because norm-responsive punishers conserve resources in defector-dominated regions while concentrating intensified sanctions at cooperative-defector boundaries, creating self-reinforcing fronts that amplify the spread of prosocial behavior. These findings reveal that enforcement efficiency can be dramatically improved by targeting punishment at cooperative-defector interfaces rather than applying uniform sanctions, offering quantitative guidelines for designing adaptive regulatory mechanisms that maximize compliance while minimizing institutional costs.

physics.soc-ph↗

arXiv2025

Adaptive network dynamics and behavioral contagion in multi-state drug use propagation

Addictive behavior spreads through social networks via feedback among choice, peer pressure, and shifting ties, a process that eludes standard epidemic models. We present a comprehensive multi-state network model that integrates utility-based behavioral transitions with adaptive network rewiring, capturing the co-evolutionary dynamics between drug use patterns and social structure. Our framework distinguishes four distinct individual states by combining drug use behavior with addiction status, while allowing individuals to strategically disconnect from drug-using neighbors and form new connections with non-users. Monte Carlo simulations show that rewiring reshapes contagion, pulling high-degree nodes into drug-free clusters and stranding users on sparse fringes. Systematic exploration of the four-dimensional parameter space reveals sharp phase transitions reminiscent of critical phenomena in statistical physics, where small changes in recovery rates or addiction conversion rates trigger dramatic shifts in population-level outcomes. Most significantly, the rewiring probability emerges as the dominant control parameter, establishing adaptive network management as more influential than biological susceptibility factors in determining addiction prevalence. Our findings challenge traditional intervention paradigms by revealing that empowering individuals to curate their social environments may be more effective than targeting individual behavioral change alone.

physics.soc-ph↗

arXiv2025

When Networks Mislead: How Partisan Communication Undermines Democratic Decision-Making

Democratic societies increasingly rely on communication networks to aggregate citizen preferences and information, yet these same networks can systematically mislead voters under certain conditions. We introduce an agent-based model that captures two rival forces in partisan networks: honest noise filtering that lifts accuracy and strategic bluffing that embeds bias. Extensive simulations show that communication architecture shapes voting accuracy more than any individual-level trait. When candidate quality gaps are moderate, partisan bluffing overpowers honest signals and steers supporters of weaker contenders into collective error. However, positioning independents in central network roles serves as an epistemic circuit breaker, preventing echo chambers from spiraling toward systematic error. Counterintuitively, we discover that competitive elections with meaningful quality differences prove most vulnerable to collective delusion. Small numbers of extreme partisans can contaminate entire communities through cascading bias effects that persist across hundreds of communication rounds. Our findings challenge conventional wisdom about information aggregation in democracy and provide actionable insights for institutional design in an era of algorithmic filtering and social media polarization.

physics.soc-ph↗

arXiv2024

Suppressing defection by increasing temptation: the impact of smart cooperators on a social dilemma situation

In a social dilemma situation, where individual and collective interests are in conflict, it sounds a reasonable assumption that the presence of super or smart players, who simultaneously punish defection and reward cooperation without allowing exploitation, could solve the basic problem. The behavior of such a multi-strategy system, however, is more subtle than it is firstly anticipated. When exploring the complete parameter space, we find that the emergence of cyclic dominance among strategies is rather common, which results in several counter-intuitive phenomena. For example, the defection level can be lowered at higher temptation, or weaker punishment provides better conditions for smart players. Our study indicates that smart cooperators can unexpectedly thrive under high temptation, emphasizing the complexity of strategic interactions. This study suggests that the principles governing these interactions can be applied to other moral behaviors, such as truth-telling and honesty, providing valuable insights for future research in multi-agent systems.

physics.soc-ph↗

arXiv2023

Supporting punishment via taxation in a structured population

Taxes are an essential and uniformly applied institution for maintaining modern societies. However, the levels of taxation remain an intensive debate topic among citizens. If each citizen contributes to common goals, a minimal tax would be sufficient to cover common expenses. However, this is only achievable at high cooperation level; hence, a larger tax bracket is required. A recent study demonstrated that if an appropriate tax partially covers the punishment of defectors, cooperation can be maintained above a critical level of the multiplication factor, characterizing the synergistic effect of common ventures. Motivated by real-life experiences, we revisited this model by assuming an interactive structure among competitors. All other model elements, including the key parameters characterizing the cost of punishment, fines, and tax level, remain unchanged. The aim was to determine how the spatiality of a population influences the competition of strategies when punishment is partly based on a uniform tax paid by all participants. This extension results in a more subtle system behavior in which different ways of coexistence can be observed, including dynamic pattern formation owing to cyclic dominance among competing strategies.

physics.soc-ph↗

arXiv2023

Restoring spatial cooperation with myopic agents in a three-strategy social dilemma

Introducing strategy complexity into the basic conflict of cooperation and defection is a natural response to avoid the tragedy of the common state. As an intermediate approach, quasi-cooperators were recently suggested to address the original problem. In this study, we test its vitality in structured populations where players have fixed partners. Naively, the latter condition should support cooperation unambiguously via enhanced network reciprocity. However, the opposite is true because the spatial structure may provide a humbler cooperation level than a well-mixed population. This unexpected behavior can be understood if we consider that at a certain parameter interval the original prisoner's dilemma game is transformed into a snow-drift game. If we replace the original imitating strategy protocol by assuming myopic players, the spatial population becomes a friendly environment for cooperation. This observation is valid in a huge region of parameter space. This study highlights that spatial structure can reveal a new aspect of social dilemmas when strategy complexity is introduced.

physics.soc-ph↗

arXiv2023

Graphlet and Orbit Computation on Heterogeneous Graphs

Many applications, ranging from natural to social sciences, rely on graphlet analysis for the intuitive and meaningful characterization of networks employing micro-level structures as building blocks. However, it has not been thoroughly explored in heterogeneous graphs, which comprise various types of nodes and edges. Finding graphlets and orbits for heterogeneous graphs is difficult because of the heterogeneity and abundance of semantic information. We consider heterogeneous graphs, which can be treated as colored graphs. By applying the canonical label technique, we determine the graph isomorphism problem with multiple states on nodes and edges. With minimal parameters, we build all non-isomorphic graphs and associated orbits. We provide a Python package that can be used to generate orbits for colored directed graphs and determine the frequency of orbit occurrence. Finally, we provide four examples to illustrate the use of the Python package.

cs.SI↗

arXiv2022

Group-size dependent synergy in heterogeneous populations

When people collaborate, they expect more in return than a simple sum of their efforts. This observation is at the heart of the so-called public goods game, where the participants' contributions are multiplied by an $r$ synergy factor before they are distributed among group members. However, a larger group could be more effective, which can be described by a larger synergy factor. To elaborate on the possible consequences, in this study, we introduce a model where the population has different sizes of groups, and the applied synergy factor depends on the size of the group. We examine different options when the increment of $r$ is linear, slow, or sudden, but in all cases, the cooperation level is higher than that in a population where the homogeneous $r$ factor is used. In the latter case, smaller groups perform better; however, this behavior is reversed when synergy increases for larger groups. Hence, the entire community benefits because larger groups are rewarded better. Notably, a similar qualitative behavior can be observed for other heterogeneous topologies, including scale-free interaction graphs.

physics.soc-ph↗

arXiv2022

When costly migration helps to improve cooperation

Motion is a typical reaction among animals and humans trying to reach better conditions in a changing world. This aspect has been studied intensively in social dilemmas where competing players' individual and collective interests are in conflict. Starting from the traditional public goods game model, where players are locally fixed and unconditional cooperators or defectors are present, we introduce two additional strategies through which agents can change their positions of dependence on the local cooperation level. More importantly, these so-called sophisticated players should bear an extra cost to maintain their permanent capacity to evaluate their neighborhood and react accordingly. Hence, four strategies compete, and the most successful one can be imitated by its neighbors. Crucially, the introduction of costly movement has a highly biased consequence on the competing main strategies. In the majority of parameter space, it is harmful to defectors and provides a significantly higher cooperation level when the population is rare. At an intermediate population density, which would be otherwise optimal for a system of immobile players, the presence of mobile actors could be detrimental if the interaction pattern changes slightly, thereby blocking the optimal percolation of information flow. In this parameter space, sophisticated cooperators can also show the co-called Moor effect by first avoiding the harmful vicinity of defectors; they subsequentially transform into an immobile cooperator state. Hence, paradoxically, the additional cost of movement could be advantageous to reach a higher general income, especially for a rare population when subgroups would be isolated otherwise.

physics.soc-ph↗

arXiv2021

Mercenary punishment in structured populations

Punishing those who refuse to participate in common efforts is a known and intensively studied way to maintain cooperation among self-interested agents. But this act is costly, hence punishers who are generally also engaged in the original joint venture, become vulnerable, which jeopardizes the effectiveness of this incentive. As an alternative, we may hire special players, whose only duty is to watch the population and punish defectors. Such a policelike or mercenary punishment can be maintained by a tax-based fund. If this tax is negligible, a cyclic dominance may emerge among different strategies. When this tax is relevant then this solution disappears. In the latter case, the fine level becomes a significant factor that determines whether punisher players coexist with cooperators or alternatively with defectors. The maximal average outcome can be reached at an intermediate cost value of punishment. Our observations highlight that we should take special care when such kind of punishment and accompanying tax are introduced to reach a collective goal.

physics.soc-ph↗

arXiv2021

Small Fraction of Selective Cooperators Can Elevate General Wellbeing Significantly

A cooperative player invests effort into a common venture without knowing the partner's intention in advance. But this strategy can be implemented in various ways when a player is involved in different games simultaneously. Interestingly, if cooperative players distinguish their neighbors and allocate all their external investments into the most successful partner's game exclusively then a significant cooperation level can be reached even at harsh circumstances where game parameters would dictate full defection otherwise. This positive impact, however, can also be reached when just a smaller fraction of players apply this sophisticated investment protocol during the game. To confirm this hypothesis we have checked several distributions that determine the fraction of supporting players who apply the mentioned selective investment protocol. Notably, when these players are not isolated, but their influences percolate then the whole population may enjoy the benefit of full cooperation already at a relatively low value of the synergy factor which represents the dilemma strength in the applied public goods game.

physics.soc-ph↗

arXiv2020

Social Contagion and Associative Diffusion in Multilayer Network

The question that how cultural variation emerges has drawn lots of interest in sociological inquiry. Sociologists predominantly study such variation through the lens of social contagion, which mostly attributes cultural variation to the underlying structural segregation, making it epiphenomenal to the pre-existing segregated structure. On the other hand, arguing culture doesn't spread like a virus, an alternative called associative diffusion was proposed, in which cultural transmission occurs not at the preference of practices, but at the association between practices. The associative diffusion model then successfully explains cultural variation without attributing it to a segregated social structure. The contagion model and associative diffusion model require different types of relationships and interactions to make cultural transmission possible. In reality, both types of relationships exist. In light of this concern, we proposed combining the two models with the multilayer network framework. On one layer, agents casually observed the behaviors of others, updating their belief about the association between practices; on another layer, agents' preference of practices are directly influenced by closed others. In the meantime, the constraint satisfaction between preference and association is used to link the update of both, thereby making each individual a coherent entity in terms of preference and association. Using this approach, we entangle the effect of social contagion and associative diffusion through multilayer networks. For the baseline, we explore the model dynamics on three common network models: fully connected, small-world, and scale-free. The results show nontrivial dynamics between the two extremes of the contagion model and the associative diffusion model, justifying our claim that it is necessary to consider the two models at the same time.

cs.SI↗

arXiv2020

Status hierarchy and group cooperation: A generalized model

In a refreshing mathematical investigation, Mark (2018) shows that status hierarchy may facilitate the emergence of cooperation in groups. Despite the contribution, the present paper notes that there are limitations in Mark's model that makes it less realistic than it could in explaining real-world experiences. Consequently, we present a more generalized modified framework in which his model is a special case, by developing and introducing a new hierarchy measure into the model to estimate the cooperation level in a set of hierarchical structures omitted in Mark's work yet common in everyday life--those with multiple leaders. We derived the conditions under which cooperation can emerge in these groups, and verified our analytical predictions in agent-based computer simulations. In so doing, not only does our model elaborate on its predecessor and support Mark's general prediction. For theory, our work further reveals two novel phenomena of group cooperation: Both the relative number of cooperators to defectors in groups and the assortativity among these different roles can backfire; they are not always the higher, the better for cooperation to thrive. For methodology, the hierarchy measure developed and our model using the measure may also be applied in future research on a wide range of related topics.

econ.GN↗

arXiv2019

Social Clustering in Epidemic Spread on Coevolving Networks

Even though transitivity is a central structural feature of social networks, its influence on epidemic spread on coevolving networks has remained relatively unexplored. Here we introduce and study an adaptive SIS epidemic model wherein the infection and network coevolve with non-trivial probability to close triangles during edge rewiring, leading to substantial reinforcement of network transitivity. This new model provides a unique opportunity to study the role of transitivity in altering the SIS dynamics on a coevolving network. Using numerical simulations and Approximate Master Equations (AME), we identify and examine a rich set of dynamical features in the new model. In many cases, the AME including transitivity reinforcement provides accurate predictions of stationary-state disease prevalences and network degree distributions. Furthermore, for some parameter settings, the AME accurately trace the temporal evolution of the system. We show that higher transitivity reinforcement in the model leads to lower levels of infective individuals in the population, when closing a triangle is the dominant rewiring mechanism. These methods and results may be useful in developing ideas and modeling strategies for controlling SIS type epidemics.

physics.soc-ph↗