Sybil-proof Answer Querying Mechanism

We study a question answering problem on a social network, where a requester is seeking an answer from the agents on the network. The goal is to design reward mechanisms to incentivize the agents to propagate the requester’s query to their neighbours if they don’t have the answer. Existing mechanisms are vulnerable to Sybil-attacks, i.e., an agent may get more reward by creating fake identities. Hence, we combat this problem by first proving some impossibility results to resolve Sybil-attacks and then characterizing a class of mechanisms which satisfy Sybil-proofness (prevents Sybil-attacks) as well as other desirable properties. Except for Sybil-proofness, we also consider cost minimization for the requester and agents’ collusions.

💡 Research Summary

This paper investigates the design of reward mechanisms for answer querying on social networks, where a requester seeks an answer from agents connected via a social network. The key challenge is to incentivize agents to propagate the query to their neighbors if they do not have the answer, while preventing Sybil attacks—where agents create fake identities to gain more rewards. The authors address this by first proving impossibility results regarding the compatibility of Sybil-proofness with other desirable properties, and then characterizing a class of mechanisms that achieve Sybil-proofness along with other incentives.

The model formalizes the query propagation process as a tree rooted at the requester, with a winner being the agent who provides the answer. The focus is on path mechanisms, which reward only agents on the winning path from the requester to the winner, with rewards depending solely on the agent’s depth and the path length. Desirable properties for the mechanism are defined: Incentive Compatibility (IC) ensures agents truthfully report answers or propagate queries; Individual Rationality (IR) and Strong IR (SIR) guarantee non-negative or positive rewards; Budget Constraint (BC) limits total reward expenditure; Sybil-Proofness (SP) prevents benefits from creating fake identities; Collusion-Proofness (CP) prevents benefits from collusion; and ρ-Split Security (ρ-SS) ensures agents receive at least a fraction ρ of their child’s reward to encourage propagation.

The paper first establishes impossibility results. Lemma 1 shows that under SIR, no path mechanism can simultaneously satisfy SP and λ-CP for all λ ≥ 3. Proposition 2 extends this to conclude that SP and CP cannot coexist under SIR. Theorem 1 characterizes that under IR, SP, and CP, the only feasible mechanisms are two-headed mechanisms, which reward only the first agent and the winner on the winning path, offering zero rewards to others. These results highlight fundamental trade-offs in mechanism design.

To overcome these limitations, the authors propose the Double Geometric Mechanism (DGM), a class of path mechanisms where rewards are determined using geometric sequences based on depth and path length. DGM is designed to satisfy IC, IR, BC, SP, 2-CP, and ρ-SS simultaneously. While SP and full CP are incompatible under SIR, DGM achieves SP and 2-CP, meaning it prevents Sybil attacks and collusions involving up to two agents. The ρ-SS property ensures that agents have sufficient incentives to propagate queries. DGM also allows the requester to control costs through budget constraints, making it practical for real-world applications.

In summary, this paper provides theoretical insights into the constraints of Sybil-proof reward mechanisms in social networks and introduces a practical mechanism that balances multiple desirable properties. By characterizing the Double Geometric Mechanism, it offers a robust solution for incentivizing query propagation while mitigating Sybil attacks and limited collusions, with implications for Q&A platforms, P2P networks, and other decentralized systems where trust and incentives are critical. The work advances the understanding of mechanism design in networked environments and sets a foundation for future research on secure and efficient query systems.