arXiv subjects

Fang-Yi Yu

Publications and source records attributed to Fang-Yi Yu.

22 records · Page 2Linked to original sources

arXiv2021

Timely Information from Prediction Markets

Prediction markets are powerful tools to elicit and aggregate beliefs from strategic agents. However, in current prediction markets, agents may exhaust the social welfare by competing to be the first to update the market. We initiate the study of the trade-off between how quickly information is aggregated by the market, and how much this information costs. We design markets to aggregate timely information from strategic agents to maximize social welfare. To this end, the market must incentivize agents to invest the correct amount of effort to acquire information: quickly enough to be useful, but not faster (and more expensively) than necessary. The market also must ensure that agents report their information truthfully and on time. We consider two settings: in the first, information is only valuable before a deadline; in the second, the value of information decreases as time passes. We use both theorems and simulations to demonstrate the mechanisms.

cs.GT↗

arXiv2020

Learning and Strongly Truthful Multi-Task Peer Prediction: A Variational Approach

Peer prediction mechanisms incentivize agents to truthfully report their signals even in the absence of verification by comparing agents' reports with those of their peers. In the detail-free multi-task setting, agents respond to multiple independent and identically distributed tasks, and the mechanism does not know the prior distribution of agents' signals. The goal is to provide an $ε$-strongly truthful mechanism where truth-telling rewards agents "strictly" more than any other strategy profile (with $ε$ additive error), and to do so while requiring as few tasks as possible. We design a family of mechanisms with a scoring function that maps a pair of reports to a score. The mechanism is strongly truthful if the scoring function is "prior ideal," and $ε$-strongly truthful as long as the scoring function is sufficiently close to the ideal one. This reduces the above mechanism design problem to a learning problem -- specifically learning an ideal scoring function. We leverage this reduction to obtain the following three results. 1) We show how to derive good bounds on the number of tasks required for different types of priors. Our reduction applies to myriad continuous signal space settings. This is the first peer-prediction mechanism on continuous signals designed for the multi-task setting. 2) We show how to turn a soft-predictor of an agent's signals (given the other agents' signals) into a mechanism. This allows the practical use of machine learning algorithms that give good results even when many agents provide noisy information. 3) For finite signal spaces, we obtain $ε$-strongly truthful mechanisms on any stochastically relevant prior, which is the maximal possible prior. In contrast, prior work only achieves a weaker notion of truthfulness (informed truthfulness) or requires stronger assumptions on the prior.

cs.GT↗

arXiv2020

Think Globally, Act Locally: On the Optimal Seeding for Nonsubmodular Influence Maximization

We study the $r$-complex contagion influence maximization problem. In the influence maximization problem, one chooses a fixed number of initial seeds in a social network to maximize the spread of their influence. In the $r$-complex contagion model, each uninfected vertex in the network becomes infected if it has at least $r$ infected neighbors. In this paper, we focus on a random graph model named the stochastic hierarchical blockmodel, which is a special case of the well-studied stochastic blockmodel. When the graph is not exceptionally sparse, in particular, when each edge appears with probability $ω(n^{-(1+1/r)})$, under certain mild assumptions, we prove that the optimal seeding strategy is to put all the seeds in a single community. This matches the intuition that in a nonsubmodular cascade model placing seeds near each other creates synergy. However, it sharply contrasts with the intuition for submodular cascade models (e.g., the independent cascade model and the linear threshold model) in which nearby seeds tend to erode each others' effects. Our key technique is a novel time-asynchronized coupling of four cascade processes. Finally, we show that this observation yields a polynomial time dynamic programming algorithm which outputs optimal seeds if each edge appears with a probability either in $ω(n^{-(1+1/r)})$ or in $o(n^{-2})$.

cs.SI↗

arXiv2020

Limitations of Greed: Influence Maximization in Undirected Networks Re-visited

We consider the influence maximization problem (selecting $k$ seeds in a network maximizing the expected total influence) on undirected graphs under the linear threshold model. On the one hand, we prove that the greedy algorithm always achieves a $(1 - (1 - 1/k)^k + Ω(1/k^3))$-approximation, showing that the greedy algorithm does slightly better on undirected graphs than the generic $(1- (1 - 1/k)^k)$ bound which also applies to directed graphs. On the other hand, we show that substantial improvement on this bound is impossible by presenting an example where the greedy algorithm can obtain at most a $(1- (1 - 1/k)^k + O(1/k^{0.2}))$ approximation. This result stands in contrast to the previous work on the independent cascade model. Like the linear threshold model, the greedy algorithm obtains a $(1-(1-1/k)^k)$-approximation on directed graphs in the independent cascade model. However, Khanna and Lucier showed that, in undirected graphs, the greedy algorithm performs substantially better: a $(1-(1-1/k)^k + c)$ approximation for constant $c > 0$. Our results show that, surprisingly, no such improvement occurs in the linear threshold model. Finally, we show that, under the linear threshold model, the approximation ratio $(1 - (1 - 1/k)^k)$ is tight if 1) the graph is directed or 2) the vertices are weighted. In other words, under either of these two settings, the greedy algorithm cannot achieve a $(1 - (1 - 1/k)^k + f(k))$-approximation for any positive function $f(k)$. The result in setting 2) is again in a sharp contrast to Khanna and Lucier's $(1 - (1 - 1/k)^k + c)$-approximation result for the independent cascade model, where the $(1 - (1 - 1/k)^k + c)$ approximation guarantee can be extended to the setting where vertices are weighted. We also discuss extensions to more generalized settings including those with edge-weighted graphs.

cs.DS↗