Juan Wang’s research while affiliated with Tianjin University of Technology and other places

... Consequently, researchers have widely incorporated Q-learning into the analysis of game dynamics to enhance cooperation in social dilemmas. [49][50][51][52][53][54][55][56][57] In summary, reinforcement learning methodologies effectively promote cooperation in game-theoretic models by necessitating that participants balance personal and collective interests through choices between cooperation and defection. However, in complex real-world contexts, individuals may prefer self-reliance over unproductive collective actions, thereby becoming "Lone Wolves" and introducing a third strategy: going it alone. ...

Reference:

Promoting cooperation in the voluntary prisoner's dilemma game via reinforcement learning

... As a further step, Li et al. 31 focused on the potential impact of second-order reputation evaluation on trust dilemmas and found that this high-order reputation evaluation helps promote prosocial behaviors to avoid the initial dilemmas, which provided a new perspective for understanding the dynamics of trust mechanisms under different social structures and contexts. Recently, Liu et al. 32 incorporated an exclusion strategy into the trust game and demonstrated that this strategy could effectively mitigate untrustworthy behavior to a certain degree, potentially even eliminating such behavior among trustees. Consequently, this approach has made a favorable impact on enhancing the overall trust level within the group. ...

Reference:

Dynamic evolution in multi-player networked trust games with graded punishment

... The development of efficient and accurate machine vision methods is crucial for enabling these robots to perform swift and precise gangue removal during raw coal transportation, thereby facilitating the transition toward environmentally sustainable coal production practices. Recent developments in convolutional neural networks have demonstrated remarkable potential in object detection tasks [8][9][10], particularly in coal-gangue discrimination applications [11]. Bounding boxes are used by object detection algorithms based on convolutional neural networks to identify the category and location of objects in images [12][13][14]. ...

Reference:

Coal and Gangue Detection Networks with Compact and High-Performance Design

... Additionally, our study mainly focuses on promoting pro-social behaviours such as trust and trustworthiness, but it would be interesting to consider the goal of optimizing social welfare in the future [58]. Furthermore, our framework assumes random interactions among group members, but in reality, individual interactions are often constrained by social networks rather than being completely random [14,[59][60][61][62]. The dynamics within the structured population deserve further investigation to offer the deeper insight into the complexity of trust evolution [63]. ...

Reference:

Evolutionary dynamics of trust in hierarchical populations with varying investment strategies

... 42 Wang et al. studied link strategy in multilayer networks and demonstrated that individuals adopting diverse behaviors could better adapt to the environment and exert a positive impact on the entire network. 43 The link strategy allows for diversified responses to different neighbors. However, without the adaptive learning capability of RL, the model still lacks flexibility and real-time strategy optimization. ...

Reference:

Cooperation resonance based on link strategy reinforcement learning and conformity

... The rapid development of complex networks, especially multi-layer networks, provides an ideal theoretical framework for characterizing the topological characteristics of these multiple complex interactions. [32][33][34][35] The powerful capabilities of multi-layer networks make them widely applicable. 36,37 For instance, multi-layer complex networks have been extended to describe many practical networks, such as aviation networks, power transmission networks, communication networks, and social networks. ...

Reference:

Third party interventions promote cooperation on the interdependent networks: A perspective based on prospect theory

... In this study [10], HDA-ResNet, a more sophisticated model for grading cervical cancer utilizing colposcopy pictures, is introduced. To capture additional image details, an architectural improvement was made by swapping out a ResNet layer with a hybrid dilated convolution layer. ...

Reference:

Classification of Cervical Cancer using ResNet-50

... In network with simplexes, we cannot neglect the case that agent i belongs to different simplexes at the same time, which means that node i may participate in the game initiated by different neighbors. On this issue, Guo et al [46] discussed the effect of the weights of payoffs that node i receives from 1-simplex and 2-simplex on the outcome, so weights of the payoffs that node i receives from 1-simplex and 2-simplex in our model are equal. Specifically, the payoff of node i can be calculated through following steps. ...

Reference:

Multiplayer game dynamics and their applications into simplicial complexes with memory effect

... PGGs are essential for understanding the mechanisms that sustain cooperation in human societies and biological populations. Several mechanisms supporting cooperation have been recognized such as direct and indirect reciprocity (cooperation can evolve when individuals interact repeatedly [HST + 18, XWPW23] or when reputation plays a role [NS98,XWPW23]), network and group structures (the spatial arrangement of individuals can promote cooperation by enabling clusters of cooperators to form and persist [SP05,SF07]) and institutional incentives (the introduction of costly punishment for defectors or rewards for contributors can sustain cooperation) [SHN01,HDP24,CSBD15,DH21b]. The latter mechanism is the focus of the present paper. ...

Reference:

Replicator-mutator dynamics for public goods games with institutional incentives

... [13][14][15][16][17] As a result, modeling and analyzing trust has become a long-standing concern of scholars from different fields and efforts to solve one of the key scientific questions. [18][19][20][21][22][23][24][25] In 1995, Berg et al. 26 proposed the theoretical modeling of trust game for the first time and quantitatively measured the trust behavior between two individuals. Subsequently, Abbass et al. 27 used the replication dynamics to analyze the evolution of trust game theoretically and proposed the evolutionary dynamics model of N-player trust game, which profoundly revealed that trust will Chaos ARTICLE pubs.aip.org/aip/cha ...

Reference:

Dynamic evolution in multi-player networked trust games with graded punishment