Exploration–exploitation tradeoff using variance estimates in multi-armed bandits

Algorithms based on upper confidence bounds for balancing exploration and exploitation are gaining popularity since they are easy to implement, efficient and effective. This paper considers a variant of the basic algorithm for the stochastic, multi-armed bandit problem that takes into account the empirical variance of the different arms. In earlier experimental works, such algorithms were found to outperform the competing algorithms. We provide the first analysis of the expected regret for such algorithms. As expected, our results show that the algorithm that uses the variance estimates has a major advantage over its alternatives that do not use such estimates provided that the variances of the payoffs of the suboptimal arms are low. We also prove that the regret concentrates only at a polynomial rate. This holds for all the upper confidence bound based algorithms and for all bandit problems except those special ones where with probability one the payoff obtained by pulling the optimal arm is larger than the expected payoff for the second best arm. Hence, although upper confidence bound bandit algorithms achieve logarithmic expected regret rates, they might not be suitable for a risk-averse decision maker. We illustrate some of the results by computer simulations.     

Recovering Multiplexing Loss through Concurrent Decode-and-Forward (DF) Relaying

In this paper, we develop a novel digital cooperative diversity transmission protocol for a two-source scenario by combining the two sources’ two classic decode-and-forward (DF) relaying steps and using 2L + 1 time slots to transmit L codewords from each source. Assuming the relays can perfectly decode their associated source messages, we give an information-theoretic average achievable capacity region for this transmission scheme. Through diversity-multiplexing tradeoff analysis, we show that our so called concurrent DF relaying protocol can effectively recover the multiplexing loss induced by the half-duplex operation in the relays, while still obtaining some diversity gain. Numerical results reveal our scheme offers significant performance advantages over the classic DF relaying protocols, especially for high signal-to-noise ratio (SNR) and large frame length L regime.

TDMAMIMO系统中利用DMT实现多用户分集度分析

针对目前多用户分集度(MUD)分析方法无法同时反映空间分集和空间复用增益的问题,提出了一种在时分多址(TDMA)多入多出系统中分析MUD的方法.首先推导了多用户中断概率;然后针对TDMA系统中给定数据速率的任意码,推导出其差错概率以多用户中断概率为下界;最后推导出该系统中给定数据速率的任意渐进通用码的差错概率上界,并由此获得最优分集复用折中(DMT)曲线.这种基于DMT的方法能在考虑空间分集度和空间复用的同时反映出MUD的性能,从而为MUD性能分析和DMT分析构建了桥梁.仿真结果表明了理论分析的正确性.     

基于分布式空时编码的异构协作通信系统介绍及DMT分析

协作通信可以在衰落无线环境中提供显著的性能增益。大多数研究协作通信的文献都假设每个节点只有一根天线或几根同样类型的天线。目前,具备异构通信接口的移动终端变得越来越普遍,所以有必要在蜂窝网络中深入地进行协作通信的相关研究。介绍了一种适用于异构协作系统的VAA-DSTC模式,并证明了其不仅可以获取满分集增益,而且移动终端异构通信接口间的数据率差异可以带来额外的复用增益。然后,重点分析了VAA-DSTC模式的分集复用折中(DMT)性能,并与其它类型的协作模式进行了比较。     

时间—资源权衡协调问题的多目标优化决策模型

时间—资源权衡协调是以追求资源消耗费用极小和项目完工时间最短为目标，在满足项目工期要求下，根据项目活动时间的先后次序与可更新资源约束有效确定项目时间表。提出了一个具有资源约束问题的时间—资源权衡协调问题的多目标优化决策数学模型。在模型中，对相互冲突的项目工期与整个被消耗的资源费用是可权衡调节的。通过权衡协调和调节项目工期与整个资源消耗费用，得到了满足权衡协调关系的满意可行解，并给出了一个数值算例。此外，在对应于资源约束的Lagrangian松弛表示式中，给出了该二人对策问题的有关特性。     

Extreme time-space tradeoffs for graphs with small space requirements

Two new isometric array languages, called Linear Array Languages (LAL) and Extended Regular Array Languages (ERAL), are introduced. A more abundant hierarchy for isometric languages is established. We show that some properties of two-dimensional array patterns do not coincide with their one-dimensional counterparts.     

Inhibition of return: Sensitivity and criterion as a function of response time.

[Correction Notice: An erratum for this article was reported in Vol 32(5) of Journal of Experimental Psychology: Human Perception and Performance (see record 2006-12344-017). On page 912, there are typographical errors in Table 1. On page 915, the last line of the left column incorrectly states that the mean response frequencies for Experiment 2 are presented within Table 2. The corrected information for both pages is presented in the erratum.] Inhibition of return (IOR) refers to a mechanism that results in a performance disadvantage typically observed when targets are presented at a location once occupied by a cue. Although the time course of the phenomenon--from the cue to the target--has been well studied, the time course of the effect--from target to response--is unknown. In 2 experiments, the effect of IOR upon sensitivity and response criterion under different levels of speed stress was examined. In go/no-go and choice reaction time tasks, IOR had at least 2 distinct effects on information processing. Early in target processing, before sufficient target information has accrued, there is a bias against responding to cued targets. Later, as target information is allowed to accrue, IOR reduces sensitivity to the target's nonspatial feature. Three accounts relating to the early bias effect of IOR and the late effect of IOR on sensitivity are offered. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

Health vs. privacy? The risk-risk tradeoff in using COVID-19 contact-tracing apps

In the midst of the COVID-19 pandemic, contact-tracing apps have emerged as reliable tools for public health communication and the promotion of preventative health. However, to function properly, contact-tracing apps require users to provide sensitive information, which has raised concerns about data disclosure, misuse and social surveillance. Little is known about how different types of risk perception simultaneously hinder and motivate individuals' engagement in mobile health apps, particularly in the context of a pandemic. Based on the privacy calculus theory and the risk-risk tradeoff concept, this study examined the risk-risk tradeoff model to enhance the understanding of COVID-19 contact-tracing app users’ decision from the perspective of risk minimization. Findings from PLS-SEM and fsQCA revealed that users engage in health risk-privacy risk tradeoff when evaluating and deciding to use the apps. The focal study therefore contributes to the research on privacy calculus theory and calls for a balanced managerial solution to mitigate this tradeoff dilemma.     

Augmenting real data with synthetic data: an application in assessing Radio‐Isotope identification algorithms

The performance of Radio‐Isotope IDentification (RIID) algorithms using gamma spectroscopy is increasingly becoming important. For example, sensors at locations that screen for illicit nuclear material rely on isotope identification to resolve innocent nuisance alarms arising from naturally occurring radioactive material. Recent data collections for RIID testing consist of repeat measurements for each of several scenarios to test RIID algorithms. Efficient allocation of measurement resources requires an appropriate number of repeats for each scenario. To help allocate measurement resources in such data collections for RIID algorithm testing, we consider using only a few real repeats per scenario. In order to reduce uncertainty in the estimated RIID algorithm performance for each scenario, the potential merit of augmenting these real repeats with realistic synthetic repeats is also considered. Our results suggest that for the scenarios and algorithms considered, approximately 10 real repeats augmented with simulated repeats will result in an estimate having comparable uncertainty to the estimate based on using 60 real repeats. Published in 2009 by John Wiley & Sons, Ltd.     

Order statistics and probabilistic robust control

Order statistics theory is applied in this paper to probabilistic robust control theory to compute the minimum sample size needed to come up with a reliable estimate of an uncertain quantity under continuity assumption of the related probability distribution. Also, the concept of distribution-free tolerance intervals is applied to estimate the range of an uncertain quantity and extract the information about its distribution. To overcome the limitations imposed by the continuity assumption in the existing order statistics theory, we have derived a cumulative distribution function of the order statistics without the continuity assumption and developed an inequality showing that this distribution has an upper bound which equals to the corresponding distribution when the continuity assumption is satisfied. By applying this inequality, we investigate the minimum computational effort needed to come up with an reliable estimate for the upper bound (or the lower bound) and the range of a quantity. We also give conditions, which are much weaker than the absolute continuity assumption, for the existence of such minimum sample size. Furthermore, the issue of making tradeoff between performance level and risk is addressed and a guideline for making this kind of tradeoff is established. This guideline can be applied in general without continuity assumption.