Multimedia Crowdsourcing With Bounded Rationality: A Cognitive Hierarchy Perspective

Abstract

In multimedia crowdsourcing, the requester's quality requirements and reward decisions will affect the workers' task selection strategies and the quality of their multimedia contributions. In this paper, we present a first study on how the workers' bounded cognitive rationality interacts with and affects the decisions and performance of a multimedia crowdsourcing system. Specifically, we consider a two-stage model, where a requester first determines the reward and the quality requirement for each task, and the workers select the tasks to accomplish accordingly. First, we consider the benchmark case where users are fully rational, and derive the requester's optimal rewards and quality requirements for the tasks. Next, we focus on the more practical bounded rational case by modeling the workers' task selection behaviors using the cognitive hierarchy theory. Comparing with the fully rational benchmark, we show that the requester can increase her profit by taking advantage of the workers' bounded cognitive rationality, especially when the workers' population is large or the workers' average cognitive level is low. When the workers' average cognitive level is very high, however, the equilibrium under the practical bounded rational model converges to that under the benchmark fully rational model. It is because workers at different levels make decisions sequentially and high cognitive level workers can accurately predict other users' strategies. Under both the fully and bounded rational models, we show that if workers are Part of this paper was presented in GLOBECOM'18 [1]. The major differences between the journal and conference versions are: a) Multimedia crowdsourcing background: In [1], we considered a general crowdsourcing platform and studied only the rewards for the tasks. In this journal submission, we further consider the quality requirement together with the reward for each task, which is important for multimedia applications. b) A more general system model: We extend our crowdsourcing model from two tasks to multiple tasks, which involves solving a challenging mixed integer non-linear programming problem. c) Performance evaluation: In this journal submission, we have some new simulation results. We demonstrate the impacts of additional parameters that were not considered in [1]. To summarize, the overall difference between the journal and conference versions exceeds 60%.

Research on the Allocation Method of Regional Science and Technology Resources from the Perspective of Rationality

“…e specific allocation methods are as follows. Based on the analysis of the application principle of rationality, the rationality perspective principle is adopted to decompose the reform path of regional science and technology resource allocation layer by layer, so that the meaning of elements at each level is complementary and cross, the upper and lower elements are parent-child relations, and the lower elements are used as evaluation indicators [15,16]. e hierarchical structure of the evaluation of the reform path of regional science and technology resource allocation is shown in Figure 1.…”

Computational Models of Human Decision-Making with Application to the Internet of Everything

“…To include the human in the system, they utilize a model based on prospect theory. Reference [9] investigates the crowdsourcing of multimedia files, where the cognitive hierarchy concept is applied to formalize the decision-making of human agents in saving and sharing files. Finally, in [10], the authors consider the cyber security problem by formulating the attacker-defender interactions as an evolutionary game.…”

“…Challenge Approach [5] File sharing Modeling human based on tag models of cooperation [6] Informative prediction of human action Optimization [7] Energy efficiency Modeling human based on data analysis [8] Repeated task allocation Modeling human utilizing prospect theory [9] Multimedia crowdsourcing Modeling human by applying cognitive hierarchy [10] Cyber security defense Modeling real-world players using quantal response dynamics as monetary compensation. Therefore, to improve efficiency, it is essential to acquire a good understanding of the utility and the behavior of the involved human decision-makers, for example, by suitably adapting the utility functions; otherwise, appropriate mechanism design is implausible.…”