Stance detection aims to assign a stance label (in favor or against) to a post towards a specific target. In the literature, there are many studies focusing on this topic, and most of them treat stance detection as a supervised learning task. Therefore, a new classifier needs to be built from scratch on a well-prepared set of ground-truth data whenever predictions are needed for an unseen target. However, it is difficult to annotate the stance of a post, since a stance is a subjective attitude towards a target. Hence, it is necessary to learn the information from unlabeled data or other target data to help stance detection with a certain target. In this study, we propose a multi-target stance detection framework to integrate multi-target data together for stance detection. Since topic and sentiment are two important factors to identify the stance of a post in multi-target data, we propose an adversarial attention network to integrate multi-target data by detecting and connecting topic and sentiment information. In particular, the adversarial network is utilized to determine the topic and the sentiment of each post to collect some target-invariant information for stance detection. In addition, the attention mechanism is utilized to connect posts with a similar topic or sentiment to acquire some key information for stance detection. The experimental results not only demonstrate the effectiveness of the proposed model, but also indicate the importance of the topic and the sentiment information for stance detection using multi-target data.

Recently, online pricing receives extensive attentions for addressing more realistic mobile crowdsensing (MCS) applications. However, existing these works mainly focus on finding an optimal price by repeated price experiments of user’s acceptance decision, which incurs a desirable regret. In practice, there are also many other factors that will affect these regrets, such contextual features as discrete features and continuous features. In this article, we investigate the impact of users’ context features and corresponding weights on selecting an optimal price. To apply for multidimensional contextual features and learned corresponding weights, three ellipsoid pricing-based mechanisms are proposed for homogenous and heterogeneous noisy sensing tasks, respectively. What is more, performance analysis are given for satisfying basic properties of crowdsensing. Finally, extensive simulations show that our mechanisms outweigh than existing mechanisms.

Recently, crowdsensing receives extensive attentions to solve complex sensing tasks. However, these existing works mainly focus on the incentive issues of traditional bidding model. In practice, solutions of the incentive issues of the posted-pricing model, are more important for some realistic mobile crowdsensing (MCS) applications. However, only few works focus on the homogeneous MCS applications (each user can perform only a single task). To this end, in this article we propose a posted-pricing incentive mechanism for heterogeneous MCS applications (different users have different limits of the number of assignments). Furthermore, considering more general MCS applications (satisfying $K$ -submodular function: the $K$ -set functions that are submodular in every type $k\in [K]$ ), we define the weighted marginal utilities and explore two posted-pricing incentive mechanisms for heterogeneous and $K$ -submodular MCS applications, respectively. Moreover, both of the two mechanisms are based not only on heterogeneous or $K$ -submodular sensing tasks, but also on heterogeneous sensing types. Finally, performance analysis shows that our mechanisms satisfy the desired properties, such as the truthfulness of acceptance decision, the individual rationality,budget feasibility,computational efficiency, and the regret minimization. Through extensive simulations, we evaluate the performance and validate the theoretical properties of our proposed mechanism.

Summary: Crowd sensing is a new paradigm that leverages pervasive sensor‐equipped mobile devices to provide sensing services like forensic analysis, documenting public spaces, and collaboratively constructing statistical models. Extensive user participation is indispensable for achieving good service quality. Nowadays, most of existing mechanisms focus on guaranteeing good service quality based on instantaneous extensive user participation for crowd sensing applications. Little attention has been dedicated to maximizing long‐term service quality for crowd sensing applications due to their heterogeneous interests, preferences, selfish behaviors, and so on. To fill these gaps, a closed‐form expression of the marginal sensing data quality based on the monopoly aggregation in economics is derived in this paper. Furthermore, we design marginal quality‐based incentive mechanisms for long‐term crowd sensing applications not only to enhance extensive user participation by maximizing the expected total profits of mobile users but also to stimulate mobile users to produce high‐quality contents by applying the marginal quality. Finally, simulation results show that our mechanisms outperform the existing solutions. Copyright © 2015 John Wiley & Sons, Ltd.

Recently, due to the proliferation and low cost of mobile devices, mobile crowdsensing (MCS) has attained extensive attentions for solving large-scale timely complex sensing tasks. Much research works mainly focus on the traditional bidding model. In practice, bidding their true cost is difficult in many real scenes. To the end, a few works investigate the posted-pricing model. However, these works bring the large regret, i.e., utility loss, due to the following issues. First, the feedback of the repeated pricing experiment is very limited, where the platform only knows whether the posted price is accepted or not. Besides, these works cannot effectively address the joint constraints of budget $B$ and deadline. According to our observation that the sensing cost of mobile users is closely relative to the contextual features, such as the changes of the sensing demands and the increases of experiences, we formalize the cost function as a linear model for heterogeneous MCS tasks. As such, we propose a context-feature budget-limited posted-pricing mechanism for heterogeneous MCS by applying the multiarm bandit decision and cost learning process. More importantly, an optimal price provided by our mechanism can be lower than an optimal price of the global fixed optimal price via the cost learning process, thereby decreasing the regret. Performance analysis demonstrates that basic properties of MCS hold in our mechanism. Moreover, our mechanism only achieves the regret $\mathcal {O}(\mathrm {ln}(B))$ when the budget is exhausted before the deadline. Extensive experiments show that our mechanism overweighs the rest of existing mechanisms.