• [Huge'20] Scalable Distributed Machine Learning with Huge Data for IoT and Scientific Discovery
  T. Luo and S. K. Das
  National Science Foundation (NSF) Large Scale Networking (LSN) Workshop on Huge Data, Chicago, IL, April 13-14, 2020.
• [AAAI'20] Adaptive Anomaly Detection for IoT Data in Hierarchical Edge Computing [pdf]
  M. V. Ngo, H. Chaouchi, T. Luo, and T. Quek
  AAAI Workshop on Artificial Intelligence of Things (AIoT), New York, NY, Feb 2020.
• [AAAI'20] COBRA: Context-aware Bernoulli neural networks for reputation assessment [pdf]
  L. Zeynalvand, T. Luo, and J. Zhang
  34th AAAI Conference on Artificial Intelligence (AAAI), New York, NY, Feb 7-12, 2020.
• [AAAI'20] Mechanism design with predicted task revenue for bike sharing systems [pdf]
  H. Lv, C. Zhang, Z. Zheng, T. Luo, F. Wu and G. Chen
  34th AAAI Conference on Artificial Intelligence (AAAI), New York, NY, Feb 7-12, 2020.
• [TOPS'20] CrowdPrivacy: Publish More Useful Data with Less Privacy Exposure in Crowdsourced Location-based Services
  F-J. Wu and T. Luo
  ACM Transactions on Privacy and Security (TOPS), 2020. To appear.
• [IoT-J'19] Improving IoT data quality in mobile crowd sensing: A cross validation approach [pdf] [DOI: 10.1109/JIOT.2019.2904704]
  T. Luo, J. Huang, S. S. Kanhere, J. Zhang, and S. K. Das
  IEEE Internet of Things Journal (IoT-J), vol. 6, no. 3, pp. 5651-5664, June 2019.
  We propose a cross validation (CV) approach which seeks a validating crowd to ratify the contributing crowd in terms of the quality of sensor data contributed by the latter. Using a weighted random oversamping (WRoS) technique and a PATOP² algorithm which makes an exploration-exploitation tradeoff, our proposed CV approach offers a unified solution to two typical yet disparate needs: reinforce obscure truth and discover hidden truth.
  (See illustration.)
• [AAIM'18] Achieving location truthfulness in rebalancing supply-demand distribution for bike sharing [pdf] [DOI: 10.1007/978-3-030-04618-7_21]
  H. Lv, F. Wu, T. Luo, X. Gao, and G. Chen
  12th International Conference on Algorithmic Aspects in Information and Management (AAIM), pp. 256-267, December 2018.
  Best Student Paper Award
• [ICC'18] Distributed anomaly detection using autoencoder neural networks in WSN for IoT [pdf] [slides] [DOI: 10.1109/ICC.2018.8422402]
  T. Luo and S. Nagarajan
  IEEE International Conference on Communications (ICC), May 2018.
  This paper is the first work that introduces autoencoder neural networks (ANN), a deep learning model, into wireless sensor networks (WSN) to detect anomalies. It contradicts the general belief that "deep learning is not suitable for WSN" by (1) "making deep learning (extremely) shallow" and (2) allocates computation load to sensors and IoT cloud using a two-part algorithm, DADA-S and DADA-C.
  (See illustration.)
• [ComMag'17] Sustainable incentives for mobile crowdsensing: Auctions, lotteries, and trust and reputation systems [pdf] [DOI: 10.1109/MCOM.2017.1600746CM]
  T. Luo, S. S. Kanhere, J. Huang, S. K. Das, and F. Wu
  IEEE Communications Magazine, vol. 55, no. 3, pp. 68-74, March 2017.
  This survey paper provides a technical overview and analysis of six incentive mechanism design frameworks: auction, lottery, trust and reputation system, bargaining game, contract theory, and market-driven mechanism.
• [TIST'16] Incentive mechanism design for crowdsourcing: an all-pay auction approach [ACM Lib] [pdf] [DOI: 10.1145/2837029]
  T. Luo, S. K. Das, H-P. Tan, and L. Xia
  ACM Transactions on Intelligent Systems and Technology (TIST), vol. 7, no. 3, pp. 35:1-26, February 2016.
  The most commonly used auctions for incentive mechanism design are winner-pay auctions, where only winners (i.e., highest bidders who will receive reward) need to pay for their bids (by money or effort). In contrast, all-pay auctions require every bidder to pay regardless of who wins, which sounds rather unreasonable. However, applying all-pay auctions to crowdsourcing, as this paper does, gains several advantages over winner-pay auctions, and reaps much higher profit with an adaptive prize.
• [TMC'16] Incentive mechanism design for heterogeneous crowdsourcing using all-pay contests [pdf] [DOI: 10.1109/TMC.2015.2485978]
  T. Luo, S. S. Kanhere, S. K. Das, and H-P. Tan
  IEEE Transactions on Mobile Computing (TMC), vol. 15, no. 9, pp. 2234-2246, September 2016.
  Despite that crowdworkers are heterogeneous in their "types" (abilities, costs, etc.) as we all know, the hardness to model and analyze it has restricted researchers to adopt a homogeneous model where all the workers' types are assumed to follow a (single) common Bayesian belief. This paper proposes an asymmetric all-pay auction model to characterize the heterogeneity, and uses a prize tuple to achieve an interesting and counter-intuitive property called Strategy Autonomy (SA).
• [INFOCOM'15] Crowdsourcing with Tullock contests: A new perspective [pdf] [DOI: 10.1109/INFOCOM.2015.7218641]
  T. Luo, S. S. Kanhere, H-P. Tan, F. Wu, and H. Wu
  The 34th IEEE International Conference on Computer Communications (INFOCOM), April 2015, pp. 2515-2523.
  Acceptance rate: 19%
  Best Paper Award nominee
  What is a Tullock contest? Think it as a lucky draw! While auctions have dominated the realm of mechanism design for decades, this paper suggests Tullock contests as an alternative mechanism that is more appealing to "ordinary" participants. Tullock contests distinguish themselves from auctions in its imperfectly discriminating property: "You always have a chance to win, no matter how 'weak' you are." This feature is particularly desirable for, e.g., micro-task crowdsourcing (such as crowd sensing).
• [TMC'15] Quality of contributed service and market equilibrium for participatory sensing [pdf] [DOI: 10.1109/TMC.2014.2330302]
  C-K. Tham and T. Luo
  IEEE Transactions on Mobile Computing (TMC), vol. 14, no. 4, pp. 829-842, April 2015.
  In order to characterize QoS for crowdsensing, this work proposes a metric called Quality of Contributed Service (QCS) which aggregates individual quality of contributions and takes into account information quality and time sensitivity. QCS is then analyzed using a market based supply-and-demand model.
• [INFOCOM'14] Profit-maximizing incentive for participatory sensing [pdf] [Much enhanced version: ACM TIST'16]
  T. Luo, H-P. Tan, and L. Xia
  The 33rd IEEE International Conference on Computer Communications (INFOCOM), April 2014, pp. 127-135.
  Acceptance rate: 19% (319 out of 1650)
• Enhancing Responsiveness and Scalability for OpenFlow Networks via Control-Message Quenching [pdf]
  T. Luo, H-P. Tan, P. C. Quan, Y-W. Law, and J. Jin
  International Conference on ICT Convergence (ICTC), October 2012, pp. 348-353.
  Best Paper Award
  

[see Complete List with instructive illustrations]