Winner Selection Research Articles

Resource sharing for collaborative edge learning: A privacy-aware incentive mechanism combined with demand prediction

Edge intelligence benefits from ubiquitous resource capabilities by numerous edge devices. However, interest concerns and privacy issues caused by the openness of the edge network make smart devices reluctant to participate in collaborative edge learning. This work proposes a privacy-aware incentive mechanism based on combinatorial double auction (PIMCDA) to facilitate idle resource sharing of diversified edge devices. First, we establish an auction model based on blockchain to form a trusted sharing market in a multi-resource binding manner. Then, we propose a two-stage auction solution combined with differential privacy: The privacy-aware winner selection stage matches the optimal winner candidate to the task that is the closest in probability to maximize the revenue of resource providers with the obfuscated location information. The probabilistic pricing decision stage is designed based on a uniform pricing method and exponential mechanism, which can ensure bid privacy from inference attacks. Furthermore, to ensure the effectiveness of PIMCDA for long-term participation, we design a supply and demand balance mechanism with a learning-based resource prediction method. Theoretical and simulation analysis demonstrate that the proposed mechanism achieves privacy-preserving in location and bid information while ensuring effective incentive properties. Our approach effectively motivates the sharing of multiple resources in edge computing.

FedAB: Truthful Federated Learning With Auction-Based Combinatorial Multi-Armed Bandit

Federated Learning (FL) emerges as a new distributed machine learning (ML) paradigm that enables thousands of mobile devices to collaboratively train ML models using local data without compromising user privacy. However, the FL learning quality highly relies on the data contribution from the distributed mobile devices. Therefore, a well-designed incentive mechanism with effectiveness, fairness, and reciprocity is in urgent need to guarantee the stable participation of users. In this paper, we propose FedAB, an incentive and client selection strategy based on a novel multi-attribute reverse auction mechanism and a combinatorial multi-armed bandit (CMAB) algorithm. First, we develop a local contribution evaluation method based on importance sampling in the FL context. We then design a novel payment mechanism that is able to preserve individual rationality and incentive compatibility (truthfulness). At last, we design a UCB-based winner selection algorithm that is proven to achieve the server’s utility maximization with fairness and reciprocity. We have conducted extensive experiments on real datasets. The results demonstrate the superiority of FedAB, with a 10-50% improvement in total reward, final accuracy, and convergence speed compared to state-of-the-art solutions.

Auction-Based Dependent Task Offloading for IoT Users in Edge Clouds

The rapid proliferation of latency-sensitive Internet of Things (IoT) applications boosts the frequency of offloading compute-intensive tasks from IoT users to mobile edge computing (MEC) due to the limitation resources of IoT devices. It is inevitably for IoT users to compete for the computing resources of the MEC, especially when the computation tasks are dependent and have hard deadline constraints. However, most existing dependent task offloading schemes may not well consider the resource competition issues among IoT users, and possibly lead to limited system performance in multiuser scenario. To address this issue, we intend to design an auction-based dependent task-offloading mechanism to improve the efficiency of task offloading for multiple IoT users. First, we formulate the dependent task offloading as a valuation maximization problem in the trade of computing resources satisfying users’ latency requirements, which has been proved to be NP-hard. Then, by jointly considering the task graph structure and the current status of the MEC, we propose a truthful auction mechanism, named greedy winner selection strategy, in which a heuristic dependent task assignment for winners is designed to improve the efficiency of the task offloading. By conducting extensive simulations, we validate that the performance of the proposed dependent task offloading strategy is superior to existing competition algorithms, in terms of total valuations, average makespans, and success rates.

Towards Privacy-Driven Truthful Incentives for Mobile Crowdsensing Under Untrusted Platform

Reverse auction-based incentive mechanisms have been commonly proposed to stimulate mobile users to participate in crowdsensing, where users submit bids to the platform to compete for interested tasks. Recent works pointed out that bid is a private information which can reveal sensitive information of users (e.g., location privacy), and proposed bidding-preserving mechanisms with differential privacy against inference attack. However, all these mechanisms rely on a trusted platform, and would fail in bid protection completely when the platform is untrusted. In this paper, we design novel privacy-preserving incentive mechanisms to protect users’ true bid information against the honest-but-curious platform while minimizing the social cost of winner selection. To this end, instead of uploading the true bid to the platform, a differentially private bid obfuscation function is designed with the exponential mechanism, which helps each user to obfuscate bids locally and submit obfuscated bids to the platform. Two solutions are proposed for the platform to solve the winner selection problem with the obfuscated information, which is proved to be NP-hard. Moreover, we further propose a novel task-bid pair protection truthful incentive mechanism to further prevent privacy leakage from the set of interested tasks, where each user encrypts his interested tasks via homomorphic encryption locally, and an encrypted task clustering method is proposed to group users with the same interested tasks into the same cluster for winner selection with users’ encrypted task-bid pairs. Both of theoretical analysis and extensive experiments demonstrate the effectiveness of proposed mechanisms against the untrusted platform.

PRAM: A Practical Sybil-Proof Auction Mechanism for Dynamic Spectrum Access With Untruthful Attackers

Auction is becoming increasingly popular for dynamic spectrum access (DSA), while it is extremely vulnerable to sybil attacks. Existing studies on sybil-proof DSA auction impractically assume that attackers bid truthfully based on true appraisals. This paper, for the first time, considers untruthful attackers and investigates the sybil-proof auction design in such more hazardous scenarios. To justify the new assumption, we first show that attackers obtain higher utilities by bidding untruthfully, especially in networks with inadequate channels. Based on this novel finding, we then design a practical sybil attack model named EqualSumBid Sybil, where attackers follow an equal-sum rule (i.e., the sum bid value of the multiple identities of an attacker equals the bid value when it bids with only one identity) instead of their true appraisals. To ensure efficient DSA under the new attack, we finally propose the PRAM, a Practical sybil-pRoof Auction Mechanism, where suspicious identity merging and bid-independent bidder sorting methods are introduced to alleviate the effect of untruthfulness on spectrum auction. Furthermore, winner selection and payment methods are designed to resist the EqualSumBid Sybil attack. Theoretical analyses and numerical results show that PRAM not only resists the EqualSumBid Sybil attack but also achieves individual rationality and truthfulness.

Truthful Incentive Mechanism for Budget-Constrained Online User Selection in Mobile Crowdsensing

Mobile crowdsensing (MCS) has attained much attention for gathering distributed mobile users to complete large-scale sensing tasks. To ensure the task completion, enough users should be motivated to participate in sensing tasks. Thus, much research in MCS focuses on proposing incentive mechanism, and these works usually focus on the offline scenario, where the information of all users is available to the platform. However, we argue that the actual MCS is usually an online scenario, where the platform does not know the user's information until they establish connections with the platform. Meanwhile, mobile users connect to the platform randomly and will cut off the connection at any time. Hence, when accessed by a user, the platform needs to make an irrevocable decision instantly about whether to select the user or not, and decides a remuneration for the user without knowing future information. In this paper, we first propose a reverse-auction framework to model the interaction between the platform and mobile users. Then, we present an online truthful incentive mechanism (OTIM) to motivate users, including online winner selection and remuneration determination strategies. Finally, massive simulations are conducted based on three real traces, and the simulation results illustrate that OTIM achieves truthfulness, individual rationality, computational efficiency and an approximately full budget utilization.

And the first runner-up is...: comparing winner selection procedures in multi-winner Tullock contests

AbstractWe characterise the strategic equivalence among k-winner contests using simultaneous and sequential winner selection. We test this prediction of strategic equivalence using a series of laboratory experiments, contrasting 1-winner contests with 2-winner contests, varying in the latter whether the outcome is revealed sequentially or in a single stage. We find that in the long run, average bidding levels are similar across strategically-equivalent contests. However, adaptation in 2-winner contests is slower and less systematic, which is consistent with the property that simultaneous winner selection results in outcomes that are more random than in the 1-winner case.

“Follower of the Reference Point”: Platform Utility-Oriented Incentive Mechanism in Crowdsensing

Crowdsensing uses the sensing units of many participants with idle resources to collect data. Since the budget of the platform is limited, it is crucial to design a mechanism to motivate participants to lower their bids. Current incentive mechanisms assume that participants’ gains and losses are absolute, but behavioral economics shows that a certain reference point determines participants’ gains and losses. Reference dependence theory shows that the reference reward given by a platform and the reward obtained before will greatly affect the decision-making of the participant. Therefore, this paper proposes a participants’ decision-making mechanism based on the reference dependence theory. We set a reference point to reduce the participants’ bids, improving the platform’s utility. At the same time, risk preference reversal theory shows that participants evaluate the benefits based on the relative value of the rewards rather than the absolute value. Therefore, this paper proposes a winner selection mechanism based on the risk preference reversal theory. Theoretical analysis and simulations demonstrate that this paper’s incentive mechanism can guarantee the platform’s utility and improve the task completion rate.

Bilateral Privacy-Preserving Truthful Incentive for Mobile Crowdsensing

Reverse auction-based incentive mechanisms have been widely adopted to stimulate mobile workers to participate in mobile crowdsensing (MCS), where workers need to provide location information for winner selection. However, most existing mechanisms rely on trusted platforms. The worker’s location data uploaded to platforms thus can be easily exposed. Recent works start to incorporate location privacy in designing incentive mechanisms, but they have not considered the task’s location privacy and the worker’s location privacy simultaneously. Therefore, we propose a bilateral location privacy-preserving mechanism for MCS on untrusted platforms. In our model, each worker adopts differential privacy to obfuscate his location locally and then submits the obfuscated location together with the bid information. Besides, instead of the exact location, the task requester is only required to upload the task profile and the task’s obfuscated location. Then, we propose the lowest-cost winner selection mechanism which aims to minimize the social cost of winner selection under the location constraint while ensuring task quality requirements, and adopt the critical payment determination mechanism to determine the payments for the winners, which satisfies truthfulness, individual rationality, and computational efficiency. Theoretical analysis and extensive experiments on real-world datasets show the effectiveness of the proposed mechanisms.

Auction-Based VM Allocation for Deadline-Sensitive Tasks in Distributed Edge Cloud

Edge cloud computing is a new paradigm in which the computation and storage services of remote cloud data centers are moved to Edge Cloud Nodes (ECNs) in network edges. Compared to traditional cloud data centers, ECNs are geographically close to mobile users so the communication latency is significantly reduced. In this paper, we study the problem of allocating Virtual Machine (VM) resources in geo-distributed ECNs to mobile users by using the auction theory. First, we treat mobile users and ECNs as the buyers and sellers of the VM resource auction, respectively. Then, we model the VM resource allocation problem as an <inline-formula><tex-math notation="LaTeX">$n$</tex-math></inline-formula> -to-one weighted bipartite graph matching problem with 0-1 knapsack constraints. Since this problem is NP-hard, we design a greedy approximation algorithm to determine the winners of the auction, based on which we propose a truthful Auction-based VM resource Allocation (AVA) mechanism to solve the problem. Moreover, we prove that the AVA mechanism not only achieves an approximately optimal solution for winner selection, but also has the properties of truthfulness, individual rationality, and computational efficiency. Finally, we conduct extensive simulations on real traces to verify the significant performances of the proposed AVA mechanism.

Victor(iou)’s myologists: snapshots of a legacy

Victor(iou)’s myologists: snapshots of a legacy

A Dynamic Rationalization of Distance Rationalizability

Distance rationalizability is an intuitive paradigm for developing and studying voting rules: given a notion of consensus and a distance function on preference profiles, a rationalizable voting rule selects an alternative that is closest to being a consensus winner. Despite its appeal, distance rationalizability faces the challenge of connecting the chosen distance measure and consensus notion to an operational measure of social desirability. We tackle this issue via the decision-theoretic framework of dynamic social choice, in which a social choice Markov decision process (MDP) models the dynamics of voter preferences in response to winner selection. We show that, for a prominent class of distance functions, one can construct a social choice MDP, with natural preference dynamics and rewards, such that a voting rule is (votewise) rationalizable with respect to the unanimity consensus for a given distance function iff it is a (deterministic) optimal policy in the MDP. This provides an alternative rationale for distance rationalizability, demonstrating the equivalence of rationalizable voting rules in a static sense and winner selection to maximize societal utility in a dynamic process.

Towards Flexible and Truthful Incentive for Multichannel Allocation in DSA

Due to its potential of increasing spectrum utilization, dynamic spectrum access (DSA) is a promising platform to solve the so-called spectrum shortage problem, caused by a huge proliferation of wireless applications and services. Because of good allocation efficiency and fairness, auction based mechanisms have been extensively studied in DSA, in which truthfulness is a critical property to guarantee that each bidder can obtain her maximal utility by bidding with her true valuation of bidding spectrum. However, existing spectrum auction mechanisms on homogenous channels strictly require bidders to bid one channel or multiple channels with continuous demands, and does not consider the noncontinuous demands for channels. In this paper, we propose THIMBLE, the first TrutHful Incentive Mechanism for flexiBle muLtichannEl allocation by considering bidders’ continuous/noncontinuous demands for channels and spatial reusability. Specifically, bidders using the same channel without interference are formulated to a group, and a group bid recalculation method is presented for noncontinuous group bids. Next, a winner selection and charging method based on group bids is designed to achieve truthfulness. Extensive simulation results show that THIMBLE achieves not only truthfulness in continuous and noncontinuous channel demand cases, but also as good spectrum allocation efficiency as existing spectrum auction mechanisms.

Crowdsourcing as a service – from pilot projects to sustainable innovation routines

While crowdsourcing may strengthen a company's innovation performance, it is only rarely embedded in organizations as an innovation practice. Our action research shows that organizations often struggle with crowdsourcing projects as they represent a different form of innovation projects and require additional capabilities and skills e.g., to frame a crowd challenge, deal with IP rights, manage the crowd, or integrate the vast input into innovation projects. To overcome these problems, organizations have to invest in project-led learning to establish easy-to-use templates and routines e.g., to handle offered incentives or the applied winner selection processes. They further need to enable business-led learning as the established innovation structures, processes, and management practices do not cope with crowdsourcing projects and present some rigidities causing high coordination efforts. Organizations that are willing to run a number of consecutive crowdsourcing projects may rapidly improve their capabilities and even come up with scalable crowdsourcing services. Our findings further suggest that crowdsourcing, digital platforms, artificial intelligence, and as-a-service approaches may also add to general project capability building.

Multi-Round Incentive Mechanism for Cold Start-Enabled Mobile Crowdsensing

Mobile CrowdSensing (MCS) has emerged as a novel paradigm for performing large-scale sensing tasks. Many incentive mechanisms have been proposed to encourage user participation in MCS. However, most of them ignore the inevitable cold start stage of MCS, where the MCS system has just begun releasing tasks. Also, they all adopt the single-round incentive without considerations of the continuous cumulative effect. Given the severe shortage of participants in the cold start stage of MCS, this paper proposes a Multi-Round Incentive Mechanism (MRIM). MRIM is based on monetary incentives by adopting multi-round cooperation and alternating between task information diffusion and task allocation operations, both of which are NP-hard problems even without inter-round coupling imposed by system budget constraints. We explore a method to predict the probability of users participating in tasks accurately. Furthermore, we present an efficient task information diffusion algorithm to maximize the number of users participating in tasks by submitting bids. We propose a fast task allocation algorithm based on truthful auction, comprising an approximation algorithm for solving the one-round winner selection and payment calculation. With budget constraints, MRIM maximizes the number of completed tasks by iteratively performing task information diffusion and task allocation. We also prove that MRIM also possesses desired properties such as computational efficiency, user rationality, platform profitability, and price truthfulness, which can further guarantee the robustness of MRIM. The extensive simulations conducted on real-world datasets have proved the efficiency of MRIM.

Optimal Favoritism in Contests with Identity-Contingent Prizes

We investigate optimal favoritism using identity-contingent prizes in a two-player Tullock model. Besides the usual balance effect, prize allocation has an extra efficiency effect: One additional unit of prize tends to induce more effort, if it is used as the winning prize for the stronger player whose marginal cost is lower. We find that a total-effort-maximizing (contest) designer should offer a larger prize to the strong player if and only if the contest is sufficiently noisy. Our results are in contrast to conventional wisdom obtained from contest models with biased winner selection rules, in which leveling the playing field is always preferable in a two-player setting.

Discovering novel memory cell designs for sentiment analysis on tweets

Designing a Recurrent Neural Network to extract sentiment from tweets is a very hard task. When using memory cells in their design, the task becomes even harder due to the large number of design alternatives and the costly process of finding a performant design. In this paper we propose an original evolutionary algorithm to address the hard challenge of discovering novel Recurrent Neural Network memory cell designs for sentiment analysis on tweets. We used three different tasks to discover and evaluate the designs. We conducted experiments and the results show that the best obtained designs surpass the baselines—which are the most popular cells, LSTM and GRU. During the discovery process we evaluated roughly 17,000 cell designs. The selected winning candidate outperformed the others for the overall sentiment analysis problem, hence showing generality. We made the winner selection by using the cumulated accuracies on all three considered tasks.

Incentive Mechanism for Mobile Devices in Dynamic Crowd Sensing System

Mobile crowdsensing (MCS) has gained much attention due to the proliferation of smart devices equipped with powerful sensors. Large-scale users are the foundation of MCS, so designing incentive mechanisms to motivate users to participate in MCS is necessary. Existing works on incentive mechanisms usually assume a scenario where a group of tasks arrive at the platform at the same time and are immediately assigned to users. We argue that a more realistic MCS scenario can delay a task, which is called the assignment duration time, to wait for appropriate users. In this scenario, we focus on proposing a truthful incentive mechanism to reduce the overall social cost. Due to the uncertainty of coming users, the problems of selecting the appropriate users and calculating the payment for each recruited user (winner) are more complicated. To overcome these challenges, we design a dynamic truthful incentive mechanism (DTIM) including winner selection and payment decision processes. The former uniformly recruits users before the assignment deadline of tasks and dynamically readjusts the recruiting frequency of other tasks to select winners iteratively, which achieves an approximation ratio. Furthermore, the latter determines truthful payment for each winner to encourage user participation as well as avoid being deceived, which achieves truthfulness, individual rationality, and computational efficiency. Finally, massive simulations based on a real dataset roma/taxi validate the DTIM, which can effectively reduce the overall social cost and make a truthful payment for each winner.

Information disclosure on the contest mechanism

In this paper, we study how to disclose the precision of the winner selection mechanism (i.e., the contest success function) that determines how effective a player’s effort is in determining his/her winning probability. Specifically, we focus on the disclosure of the adopted discriminatory power r of the Tullock contest. The discriminatory power r is exogenously given, and the contest organizer knows more than contestants about the true state of r. We examine the optimal disclosure policy that maximizes the expected sum of contestants’ efforts. In addition to public disclosure, we study private disclosure over groups with independent signals. We establish the optimality of full disclosure/concealment.

Effect of Crowd Voting on Participation in Crowdsourcing Contests

ABSTRACT While expert rating is still a dominant approach for selecting winners in contests for creative works, a few crowdsourcing platforms have recently used “crowd voting” for winner selection – that is, let users of the crowdsourcing community publicly vote for contest winners. We investigate how a contest’s reliance on crowd voting for winner selection, defined as the percentage of crowd-voted prizes to the total prize sum (in dollar amounts), affects contest participation. Drawing upon expectancy theory and tournament theory, we develop a theoretical understanding of this relationship. Using a novel dataset of contests employing both crowd voting and expert rating, we find that a contest’s reliance on crowd voting is positively associated with participation. Specifically, every 10% increase in the crowd-voting reliance can boost users’ odds of participation by about 7%. Moreover, crowd voting is more appealing to users whose expertise is not high and whose status in the crowdsourcing community is high.