@inproceedings{wang_augmenting_2022,
	title = {Augmenting {Knowledge} {Graphs} for {Better} {Link} {Prediction}},
	volume = {3},
	url = {https://www.ijcai.org/proceedings/2022/316},
	doi = {10.24963/ijcai.2022/316},
	abstract = {Electronic proceedings of IJCAI 2022},
	language = {en},
	urldate = {2023-02-18},
	author = {Wang, Jiang and Ilievski, Filip and Szekely, Pedro and Yao, Ke-Thia},
	month = jul,
	year = {2022},
	note = {ISSN: 1045-0823},
	pages = {2277--2283},
	file = {Full Text PDF:C\:\\Users\\ktyao\\Zotero\\storage\\BFVZ3WTJ\\Wang et al. - 2022 - Augmenting Knowledge Graphs for Better Link Predic.pdf:application/pdf},
}

Electronic proceedings of IJCAI 2022

@inproceedings{yao_improved_2022,
	title = {Improved {Operational} {Carbon} {Footprint} {Reduction} by {Using} a {Data}-{Driven} {Method} in {Chemical} {Treatment} {Candidate} {Selection}},
	url = {https://onepetro.org/SPEATCE/proceedings-abstract/22ATCE/3-22ATCE/509047},
	doi = {10.2118/209999-MS},
	abstract = {Abstract. This paper presents a physics based and data driven machine learning approach for chemical treatment candidate well selection in fields producing heavy oil. In heavy oil fields, cyclic steam is often used to not only to stimulate the formation around a wellbore but also to open the casing perforations or liner slots plugged by asphaltene deposits. These asphaltene and paraffin build ups can partially or fully block the passages to fluid production. Such cyclic steam jobs are done when there is significant decline in production as measured with usually a threshold over 60 days of operation. But generating steam and using it in cyclic steam jobs is costly and contribute to the carbon footprint. An alternative solution is to employ chemical treatment of wells to clean up blockages that may occur in formations (Brownlee and Zern, 1997). When there is true blockage due to deposits by heavier oil components, cleaning the perforations or liner slots is expected to improve flow and increase the production. However, historical data show that not all chemical treatment jobs yield the same level of success. The problem addressed in this paper is using a data-driven predictive modeling approach to make recommendations for candidate well selection using chemical treatment based on past performance of chemically treated wells. This paper presents the development of a fully customized machine learning tool that has been designed as a decision support tool for a particular task for non-data scientist reservoir and production engineers. The architecture of this tool includes independent modules that can be re-trained in the future once additional production data becomes available from the fields. We attribute the success of this research to the fact that there was active collaboration among domain experts. Critical physical properties were included in this study, including chemistry studies of chemicals and limited field trials in validating the developed model. The resulting machine learning model can be used by field engineers for planning chemical well treatments given the continuously updated past performance data.},
	language = {en},
	urldate = {2023-02-18},
	publisher = {OnePetro},
	author = {Yao, Ke-Thia and Raghavenda, Cauligi S. and Ashayeri, Cyrus and Trivedi, Dhruvil and Patel, Tirth and Lal, Manish and Row, Richard and Aksehirli, Attila and Ershaghi, Iraj},
	month = sep,
	year = {2022},
	file = {Full Text PDF:C\:\\Users\\ktyao\\Zotero\\storage\\QJJGVHKB\\Yao et al. - 2022 - Improved Operational Carbon Footprint Reduction by.pdf:application/pdf},
}

Abstract. This paper presents a physics based and data driven machine learning approach for chemical treatment candidate well selection in fields producing heavy oil. In heavy oil fields, cyclic steam is often used to not only to stimulate the formation around a wellbore but also to open the casing perforations or liner slots plugged by asphaltene deposits. These asphaltene and paraffin build ups can partially or fully block the passages to fluid production. Such cyclic steam jobs are done when there is significant decline in production as measured with usually a threshold over 60 days of operation. But generating steam and using it in cyclic steam jobs is costly and contribute to the carbon footprint. An alternative solution is to employ chemical treatment of wells to clean up blockages that may occur in formations (Brownlee and Zern, 1997). When there is true blockage due to deposits by heavier oil components, cleaning the perforations or liner slots is expected to improve flow and increase the production. However, historical data show that not all chemical treatment jobs yield the same level of success. The problem addressed in this paper is using a data-driven predictive modeling approach to make recommendations for candidate well selection using chemical treatment based on past performance of chemically treated wells. This paper presents the development of a fully customized machine learning tool that has been designed as a decision support tool for a particular task for non-data scientist reservoir and production engineers. The architecture of this tool includes independent modules that can be re-trained in the future once additional production data becomes available from the fields. We attribute the success of this research to the fact that there was active collaboration among domain experts. Critical physical properties were included in this study, including chemistry studies of chemicals and limited field trials in validating the developed model. The resulting machine learning model can be used by field engineers for planning chemical well treatments given the continuously updated past performance data.

@article{liu_boltzmann_2020,
	title = {Boltzmann machine modeling of layered {MoS2} synthesis on a quantum annealer},
	volume = {173},
	issn = {0927-0256},
	url = {http://www.sciencedirect.com/science/article/pii/S0927025619307281},
	doi = {10.1016/j.commatsci.2019.109429},
	abstract = {Two-dimensional layered materials are usually synthesized by chemical vapor deposition (CVD). Here we use the Restricted Boltzmann Machine (RBM) and its enhancement called Limited Boltzmann Machine (LBM) on a quantum computer to model the chemical vapor deposition (CVD) growth data for a MoS2 monolayer. The CVD data for Boltzmann machines are generated by reactive molecular dynamics (RMD) simulations. The Boltzmann machines give insight into the semiconducting (2H) and metallic (1T) phases and defects generated during the CVD growth of MoS2. We compare the performances of the two types of Boltzmann machine models, i.e. RBM versus augmented topologies with limited intra-layer coupling between the hidden units (LBM). We show that limited topologies have a performance advantage over restricted topologies. We examine connectivity within our topological variants, explore hardware qubit mapping schemes, and discuss what performance differences may imply about locality within the data without prior knowledge.},
	language = {en},
	urldate = {2020-10-05},
	journal = {Computational Materials Science},
	author = {Liu, Jeremy and Mohan, Ankith and Kalia, Rajiv K. and Nakano, Aiichiro and Nomura, Ken-ichi and Vashishta, Priya and Yao, Ke-Thia},
	month = feb,
	year = {2020},
	keywords = {Boltzmann machine, Layered material, Quantum annealer},
	pages = {109429},
	file = {ScienceDirect Full Text PDF:C\:\\Users\\ktyao\\Zotero\\storage\\QS6MKFYD\\Liu et al. - 2020 - Boltzmann machine modeling of layered MoS2 synthes.pdf:application/pdf;ScienceDirect Snapshot:C\:\\Users\\ktyao\\Zotero\\storage\\CIRJZGGG\\S0927025619307281.html:text/html},
}

Two-dimensional layered materials are usually synthesized by chemical vapor deposition (CVD). Here we use the Restricted Boltzmann Machine (RBM) and its enhancement called Limited Boltzmann Machine (LBM) on a quantum computer to model the chemical vapor deposition (CVD) growth data for a MoS2 monolayer. The CVD data for Boltzmann machines are generated by reactive molecular dynamics (RMD) simulations. The Boltzmann machines give insight into the semiconducting (2H) and metallic (1T) phases and defects generated during the CVD growth of MoS2. We compare the performances of the two types of Boltzmann machine models, i.e. RBM versus augmented topologies with limited intra-layer coupling between the hidden units (LBM). We show that limited topologies have a performance advantage over restricted topologies. We examine connectivity within our topological variants, explore hardware qubit mapping schemes, and discuss what performance differences may imply about locality within the data without prior knowledge.

@article{liu_dynamic_2020,
	title = {Dynamic {Topology} {Reconfiguration} of {Boltzmann} {Machines} on {Quantum} {Annealers}},
	volume = {22},
	copyright = {http://creativecommons.org/licenses/by/3.0/},
	url = {https://www.mdpi.com/1099-4300/22/11/1202},
	doi = {10.3390/e22111202},
	abstract = {Boltzmann machines have useful roles in deep learning applications, such as generative data modeling, initializing weights for other types of networks, or extracting efficient representations from high-dimensional data. Most Boltzmann machines use restricted topologies that exclude looping connectivity, as such connectivity creates complex distributions that are difficult to sample. We have used an open-system quantum annealer to sample from complex distributions and implement Boltzmann machines with looping connectivity. Further, we have created policies mapping Boltzmann machine variables to the quantum bits of an annealer. These policies, based on correlation and entropy metrics, dynamically reconfigure the topology of Boltzmann machines during training and improve performance.},
	language = {en},
	number = {11},
	urldate = {2021-02-01},
	journal = {Entropy},
	author = {Liu, Jeremy and Yao, Ke-Thia and Spedalieri, Federico},
	month = nov,
	year = {2020},
	note = {Number: 11
Publisher: Multidisciplinary Digital Publishing Institute},
	keywords = {Boltzmann machines, machine learning, algorithms, entropy, quantum annealing},
	pages = {1202},
	file = {Full Text PDF:C\:\\Users\\ktyao\\Zotero\\storage\\IYQGTX3R\\Liu et al. - 2020 - Dynamic Topology Reconfiguration of Boltzmann Mach.pdf:application/pdf;Snapshot:C\:\\Users\\ktyao\\Zotero\\storage\\ZVDCRXNM\\1202.html:text/html},
}

Boltzmann machines have useful roles in deep learning applications, such as generative data modeling, initializing weights for other types of networks, or extracting efficient representations from high-dimensional data. Most Boltzmann machines use restricted topologies that exclude looping connectivity, as such connectivity creates complex distributions that are difficult to sample. We have used an open-system quantum annealer to sample from complex distributions and implement Boltzmann machines with looping connectivity. Further, we have created policies mapping Boltzmann machine variables to the quantum bits of an annealer. These policies, based on correlation and entropy metrics, dynamically reconfigure the topology of Boltzmann machines during training and improve performance.

@inproceedings{garijo_template-based_2020,
	series = {{CEUR} {Workshop} {Proceedings}},
	title = {A {Template}-{Based} {Approach} for {Annotating} {Long}-{Tail} {Datasets}},
	volume = {2722},
	url = {http://ceur-ws.org/Vol-2722/profiles2020-paper-3.pdf},
	booktitle = {Joint {Proceedings} of {Workshops} {AI4LEGAL2020}, {NLIWOD}, {PROFILES} 2020, {QuWeDa} 2020 and {SEMIFORM2020} {Colocated} with the 19th {International} {Semantic} {Web} {Conference} ({ISWC} 2020), {Virtual} {Conference}, {November}, 2020},
	publisher = {CEUR-WS.org},
	author = {Garijo, Daniel and Yao, Ke-Thia and Singh, Amandeep and Szekely, Pedro A.},
	editor = {Koubarakis, Manolis and Alani, Harith and Antoniou, Grigoris and Bontcheva, Kalina and Breslin, John G. and Collarana, Diego and Demidova, Elena and Dietze, Stefan and Gottschalk, Simon and Governatori, Guido and Hogan, Aidan and Lécué, Freddy and Montiel-Ponsoda, Elena and Ngomo, Axel-Cyrille Ngonga and Pinto, Sofia and Saleem, Muhammad and Troncy, Raphaël and Tsalapati, Eleni and Usbeck, Ricardo and Verborgh, Ruben},
	year = {2020},
	pages = {97--102},
}

@article{liu_adiabatic_2018,
	title = {Adiabatic {Quantum} {Computation} {Applied} to {Deep} {Learning} {Networks}},
	volume = {20},
	url = {https://doi.org/10.3390/e20050380},
	doi = {10.3390/e20050380},
	number = {5},
	journal = {Entropy},
	author = {Liu, Jeremy and Spedalieri, Federico M. and Yao, Ke-Thia and Potok, Thomas E. and Schuman, Catherine D. and Young, Steven R. and Patton, Robert M. and Rose, Garrett S. and Chakma, Gangotree},
	year = {2018},
	pages = {380},
	file = {Liu et al. - 2018 - Adiabatic Quantum Computation Applied to Deep Lear.pdf:C\:\\Users\\ktyao\\Zotero\\storage\\V9TW3J8J\\Liu et al. - 2018 - Adiabatic Quantum Computation Applied to Deep Lear.pdf:application/pdf},
}

@article{potok_study_2018,
	title = {A {Study} of {Complex} {Deep} {Learning} {Networks} on {High}-{Performance}, {Neuromorphic}, and {Quantum} {Computers}},
	volume = {14},
	url = {https://doi.org/10.1145/3178454},
	doi = {10.1145/3178454},
	number = {2},
	journal = {JETC},
	author = {Potok, Thomas E. and Schuman, Catherine D. and Young, Steven R. and Patton, Robert M. and Spedalieri, Federico M. and Liu, Jeremy and Yao, Ke-Thia and Rose, Garrett S. and Chakma, Gangotree},
	year = {2018},
	pages = {19:1--19:21},
}

@inproceedings{gil_p4ml:_2018,
	title = {{P4ML}: {A} {Phased} {Performance}-{Based} {Pipeline} {Planner} for {Automated} {Machine} {Learning}},
	url = {http://www.isi.edu/ gil/papers/gil-etal-automl18.pdf},
	booktitle = {Proceedings of {Machine} {Learning} {Research}, {ICML} 2018 {AutoML} {Workshop}},
	author = {Gil, Yolanda and Yao, Ke-Thia and Ratnakar, Varun and Garijo, Daniel and Steeg, Greg Ver and Szekely, Pedro and Brekelmans, Rob and Kejriwal, Mayank and Luo, Fanghao and Huang, I.-Hui},
	year = {2018},
}

@inproceedings{liu_characterizing_2017,
	title = {Characterizing {Molecular}-{Dynamics} {Simulations} {Using} {Non}-{Local} {Spatial}-{Temporal} {Metrics}},
	abstract = {Improving the quality of molecular dynamics simulations requires optimization, and the choice of objectives heavily impacts parameter search. Current work with reactive molecular dynamics (RMD) uses simple atomic bond counts as optimization targets, but these metrics have weaknesses. Here we propose non-local spatial-temporal approaches for describing simulation states and for characterizing simulation quality as compared to the ground truth. Non-local metrics look beyond individual atoms and consider molecular structure when describing a simulation state. Spatial-temporal metrics measure how the simulation state evolves over time instead of focusing solely on the current timestep. Using these new metrics in optimization may enable generation of higher quality simulations.},
	booktitle = {Proceedings of the 2017 {International} {Conference} on {Scientific} {Computing}},
	author = {Liu, Jeremy and Yao, Ke-Thia and Nakano, Aiichiro and Rajak, Pankaj},
	year = {2017},
	pages = {155--158},
}

Improving the quality of molecular dynamics simulations requires optimization, and the choice of objectives heavily impacts parameter search. Current work with reactive molecular dynamics (RMD) uses simple atomic bond counts as optimization targets, but these metrics have weaknesses. Here we propose non-local spatial-temporal approaches for describing simulation states and for characterizing simulation quality as compared to the ground truth. Non-local metrics look beyond individual atoms and consider molecular structure when describing a simulation state. Spatial-temporal metrics measure how the simulation state evolves over time instead of focusing solely on the current timestep. Using these new metrics in optimization may enable generation of higher quality simulations.

@inproceedings{potok_study_2016,
	address = {Piscataway, NJ, USA},
	series = {{MLHPC} '16},
	title = {A {Study} of {Complex} {Deep} {Learning} {Networks} on {High} {Performance}, {Neuromorphic}, and {Quantum} {Computers}},
	isbn = {978-1-5090-3882-4},
	url = {https://doi.org/10.1109/MLHPC.2016.9},
	doi = {10.1109/MLHPC.2016.9},
	abstract = {Current Deep Learning models use highly optimized convolutional neural networks (CNN) trained on large graphical processing units (GPU)-based computers with a fairly simple layered network topology, i.e., highly connected layers, without intra-layer connections. Complex topologies have been proposed, but are intractable to train on current systems. Building the topologies of the deep learning network requires hand tuning, and implementing the network in hardware is expensive in both cost and power. In this paper, we evaluate deep learning models using three different computing architectures to address these problems: quantum computing to train complex topologies, high performance computing (HPC) to automatically determine network topology, and neuromorphic computing for a low-power hardware implementation. Due to input size limitations of current quantum computers we use the MNIST dataset for our evaluation. The results show the possibility of using the three architectures in tandem to explore complex deep learning networks that are untrainable using a von Neumann architecture. We show that a quantum computer can find high quality values of intra-layer connections and weights, while yielding a tractable time result as the complexity of the network increases; a high performance computer can find optimal layer-based topologies; and a neuromorphic computer can represent the complex topology and weights derived from the other architectures in low power memristive hardware. This represents a new capability that is not feasible with current von Neumann architecture. It potentially enables the ability to solve very complicated problems unsolvable with current computing technologies.},
	urldate = {2017-01-13},
	booktitle = {Proceedings of the {Workshop} on {Machine} {Learning} in {High} {Performance} {Computing} {Environments}},
	publisher = {IEEE Press},
	author = {Potok, Thomas E. and Schuman, Catherine D. and Young, Steven R. and Patton, Robert M. and Spedalieri, Federico and Liu, Jeremy and Yao, Ke-Thia and Rose, Garrett and Chakma, Gangotree},
	year = {2016},
	pages = {47--55},
	file = {ACM Full Text PDF:C\:\\Users\\ktyao\\Zotero\\storage\\ZG6M9PZ8\\Potok et al. - 2016 - A Study of Complex Deep Learning Networks on High .pdf:application/pdf},
}

Current Deep Learning models use highly optimized convolutional neural networks (CNN) trained on large graphical processing units (GPU)-based computers with a fairly simple layered network topology, i.e., highly connected layers, without intra-layer connections. Complex topologies have been proposed, but are intractable to train on current systems. Building the topologies of the deep learning network requires hand tuning, and implementing the network in hardware is expensive in both cost and power. In this paper, we evaluate deep learning models using three different computing architectures to address these problems: quantum computing to train complex topologies, high performance computing (HPC) to automatically determine network topology, and neuromorphic computing for a low-power hardware implementation. Due to input size limitations of current quantum computers we use the MNIST dataset for our evaluation. The results show the possibility of using the three architectures in tandem to explore complex deep learning networks that are untrainable using a von Neumann architecture. We show that a quantum computer can find high quality values of intra-layer connections and weights, while yielding a tractable time result as the complexity of the network increases; a high performance computer can find optimal layer-based topologies; and a neuromorphic computer can represent the complex topology and weights derived from the other architectures in low power memristive hardware. This represents a new capability that is not feasible with current von Neumann architecture. It potentially enables the ability to solve very complicated problems unsolvable with current computing technologies.

@article{yao_system_2016,
	title = {System {Engineering} a {Scalable} {Cyber} {Warfare} {Test} {Environment}},
	volume = {37},
	number = {1},
	journal = {The ITEA Journal},
	author = {Yao, Ke-Thia and Tran, John J. and Davis, Dan M. and Burns, Daniel P. and Lucas, Robert F.},
	month = mar,
	year = {2016},
}

@inproceedings{castello_employing_2016,
	title = {Employing {High} {Performance} {Computing} to {Realize} a {Cyber} {Quick}-{Reaction} {Training} {Environment}},
	booktitle = {Modsim {World}},
	author = {Castello, Brian and Tran, John and Hire, Douglas and Lucas, Robert and Yao, Ke-Thia},
	year = {2016},
}

@inproceedings{sharifi_prediction_2015,
	title = {Prediction of {Remaining} {Life} in {Pipes} using {Machine} {Learning} from {Thickness} {Measurements}},
	booktitle = {{SPE}-173995-{MS}},
	publisher = {Society of PetroleumEngineers},
	author = {Sharifi, Mona and Yao, Ke-Thia and Raghavendra, Cauligi S. and Ershaghi, Iraj and House, Robert and Blouin, Jacques},
	month = apr,
	year = {2015},
}

@inproceedings{guo_data_2015,
	title = {Data {Driven} {Approach} to {Failure} {Prediction} for {Electrical} {Submersible} {Pump} {Systems}},
	booktitle = {{SPE}-174062-{MS}},
	publisher = {Society of Petroleum Engineers},
	author = {Guo, Dong and Raghavendra, Cauligi S. and Yao, Ke-Thia and Harding, Mark and Anvar, Amir and Patel, Anil},
	month = apr,
	year = {2015},
}

@inproceedings{liu_autoencoder-derived_2015,
	title = {Autoencoder-derived {Features} as {Inputs} to {Classification} {Algorithms} for {Predicting} {Well} {Failures}},
	booktitle = {{SPE}-174015-{MS}},
	publisher = {Society of Petroleum Engineers},
	author = {Liu, Jeremy and Jaiswal, Ayush and Yao, Ke-Thia and Raghavendra, Cauligi S.},
	month = apr,
	year = {2015},
}

@inproceedings{liu_global_2013,
	title = {Global {Model} for {Failure} {Prediction} for {Rod} {Pump} {Artificial} {Lift} {Systems}},
	booktitle = {{SPE} {Western} {Regional} {Conference}},
	publisher = {Society of Petroleum Engineers},
	author = {Liu, Yintao and Yao, Ke-Thia and Raghavendra, Cauligi S. and Wu, Anqi and Guo, Dong and Zheng, Jingwen and Olabinjo, Lanre and Balogun, Oluwafemi and Ershaghi, Iraj},
	year = {2013},
}

@inproceedings{liu_weighted_2013,
	title = {Weighted {Task} {Regularization} for {Multitask} {Learning}},
	doi = {10.1109/ICDMW.2013.158},
	abstract = {Multitask Learning has been proven to be more effective than the traditional single task learning on many real-world problems by simultaneously transferring knowledge among different tasks which may suffer from limited labeled data. However, in order to build a reliable multitask learning model, nontrivial effort to construct the relatedness between different tasks is critical. When the number of tasks is not large, the learning outcome may suffer if there exists outlier tasks that inappropriately bias majority. Rather than identifying or discarding such outlier tasks, we present a weighted regularized multitask learning framework based on regularized multitask learning, which uses statistical metrics, such as Kullback-Leibler divergence, to assign weights prior to regularization process that robustly reduces the impact of outlier tasks and results in better learned models for all tasks. We then show that this formulation can be solved using dual form like optimizing a standard support vector machine with varied kernels. We perform experiments using both synthetic dataset and real-world dataset from petroleum industry which shows that our methodology outperforms existing methods.},
	booktitle = {2013 {IEEE} 13th {International} {Conference} on {Data} {Mining} {Workshops}},
	author = {Liu, Yintao and Wu, Anqi and Guo, Dong and Yao, Ke-Thia and Raghavendra, Cauligi S.},
	month = dec,
	year = {2013},
	note = {ISSN: 2375-9259},
	keywords = {Kernel, support vector machines, Training, Accuracy, Optimization, Support vector machines, anomaly detection, Educational institutions, Equations, support vector machine, learning (artificial intelligence), data handling, knowledge transfer, Kullback-Leibler divergence, multitask learning, outlier task, regularization process, statistical metrics, svm, weighted regularization, weighted regularized multitask learning framework, weighted task regularization},
	pages = {399--406},
	file = {IEEE Xplore Abstract Record:C\:\\Users\\ktyao\\Zotero\\storage\\TY8LIMP4\\6753948.html:text/html},
}

Multitask Learning has been proven to be more effective than the traditional single task learning on many real-world problems by simultaneously transferring knowledge among different tasks which may suffer from limited labeled data. However, in order to build a reliable multitask learning model, nontrivial effort to construct the relatedness between different tasks is critical. When the number of tasks is not large, the learning outcome may suffer if there exists outlier tasks that inappropriately bias majority. Rather than identifying or discarding such outlier tasks, we present a weighted regularized multitask learning framework based on regularized multitask learning, which uses statistical metrics, such as Kullback-Leibler divergence, to assign weights prior to regularization process that robustly reduces the impact of outlier tasks and results in better learned models for all tasks. We then show that this formulation can be solved using dual form like optimizing a standard support vector machine with varied kernels. We perform experiments using both synthetic dataset and real-world dataset from petroleum industry which shows that our methodology outperforms existing methods.

@inproceedings{liu_automatic_2011,
	title = {Automatic {Early} {Failure} {Detection} for {Rod} {Pump} {Systems}},
	booktitle = {Annual {Technical} {Conference} and {Exhibition} ({ATCE} 2011)},
	publisher = {Society of Petroleum Engineers (SPE)},
	author = {Liu, Shuping and Raghavendra, Cauligi S. and Liu, Yintao and Yao, Ke-Thia and Lenz, Tracy L. and Olabinjo, Lanre and Balogun, Oluwafemi and Babu, C. G. Dinesh and Ershaghi, Iraj},
	year = {2011},
}

@inproceedings{liu_semi-supervised_2011,
	title = {Semi-supervised {Failure} {Prediction} for {Oil} {Production} {Wells}},
	doi = {10.1109/ICDMW.2011.151},
	abstract = {In the petroleum industry, multivariate time series data is commonly used to monitor the performance of their assets, in which wells artificial lift systems are among the key assets that bring oil up to the surface. Failures frequently occur among these artificial lift systems, and they can greatly increase the operational expense due to loss of production and cost of repairs (also known as workovers). Predicting these failures before they occur can dramatically improve operational performance, such as by adjusting operating parameters to forestall failures or by scheduling maintenance to reduce unplanned repairs and to minimize downtime. Artificial lift failure prediction problem poses interesting challenges to data mining algorithms, because of the many real-world data issues, such as noise, missing data, delay of failure event logs, and large variability among normally functioning well artificial lift units. This paper presents the Smart Engineering Apprentice (SEA) framework that incorporates robust feature extraction algorithm, clustering and semi-supervised learning techniques, to enable learning of failure/normal patterns from noisy and poorly labeled multivariate time series, while achieving a high recall and precision for failures for real-world dataset.},
	booktitle = {2011 {IEEE} 11th {International} {Conference} on {Data} {Mining} {Workshops}},
	author = {Liu, Yintao and Yao, Ke-Thia and Liu, Shuping and Raghavendra, Cauligi S. and Balogun, Oluwafemi and Olabinjo, Lanre},
	month = dec,
	year = {2011},
	note = {ISSN: 2375-9259},
	keywords = {Feature extraction, feature extraction, Hidden Markov models, scheduling, semi-supervised learning, artificial lift systems, asset performance monitoring, clustering, condition monitoring, data mining, Data mining, data mining algorithms, downtime minimization, failure analysis, failure prediction, hydrocarbon reservoirs, Labeling, lifts, maintenance engineering, Maintenance engineering, maintenance scheduling, multiple multivariate time series, multivariate time series data, oil production wells, petroleum, petroleum industry, Production, robust feature extraction algorithm, semi supervised failure prediction, smart engineering apprentice framework, time series, Time series analysis, unplanned repair reduction},
	pages = {434--441},
	file = {IEEE Xplore Abstract Record:C\:\\Users\\ktyao\\Zotero\\storage\\ZRZVC4RD\\6137412.html:text/html},
}

In the petroleum industry, multivariate time series data is commonly used to monitor the performance of their assets, in which wells artificial lift systems are among the key assets that bring oil up to the surface. Failures frequently occur among these artificial lift systems, and they can greatly increase the operational expense due to loss of production and cost of repairs (also known as workovers). Predicting these failures before they occur can dramatically improve operational performance, such as by adjusting operating parameters to forestall failures or by scheduling maintenance to reduce unplanned repairs and to minimize downtime. Artificial lift failure prediction problem poses interesting challenges to data mining algorithms, because of the many real-world data issues, such as noise, missing data, delay of failure event logs, and large variability among normally functioning well artificial lift units. This paper presents the Smart Engineering Apprentice (SEA) framework that incorporates robust feature extraction algorithm, clustering and semi-supervised learning techniques, to enable learning of failure/normal patterns from noisy and poorly labeled multivariate time series, while achieving a high recall and precision for failures for real-world dataset.

@inproceedings{lee_dynamic_2010,
	title = {Dynamic structure {Learning} of {Factor} {Graphs} and {Parameter} {Estimation} of a {Constrained} {Nonlinear} {Predictive} {Model} for {Oilfield} {Optimization}},
	booktitle = {Proceedings of the 2010 {International} {Conference} on {Artificial} {Intelligence}},
	author = {Lee, Hyokyeong and Yao, Ke-Thia and Nakano, Aiichiro},
	month = jul,
	year = {2010},
}

@inproceedings{lee_identifying_2010,
	title = {Identifying {Injector}-{Producer} {Relationship} in {Waterflood} {Using} {Hybrid} {Constrained} {Nonlinear} {Optimization}},
	booktitle = {{SPE} {Western} {Regional} {Meeting}},
	author = {Lee, Hyokyeong and Yao, Ke-Thia and Okpani, Olu Ogbonnaya and Nakano, Aiichiro and Ershaghi, Iraj},
	month = may,
	year = {2010},
	note = {SPE 132359.},
}

@article{lee_hybrid_2010,
	title = {Hybrid {Constrained} {Nonlinear} {Optimization} to {Injector}-{Producer} {Relationships} in {Oil} {Fields}},
	journal = {International Journal of Computer Science},
	author = {Lee, Hyokyeong and Yao, Ke-Thia and Okpani, Olu Ogbonnaya and Nakano, Aiichiro and Ershaghi, Iraj},
	year = {2010},
}

@inproceedings{liu_failure_2010,
	title = {Failure {Prediction} for {Artificial} {Lift} {Systems}},
	booktitle = {{SPE} {Western} {Regional} {Meeting}},
	author = {Liu, Yintao and Yao, Ke-Thia and Liu, Shuping and Raghavendra, Cauligi S. and Lenz, Tracy L. and Olabinjo, Lanre and Seren, Burcu and Seddighrad, Sanaz and Babu, Chinnapparaja Gunaskaran Dinesh},
	month = may,
	year = {2010},
	note = {SPE 133545.},
}

@inproceedings{gottschalk_distributed_2010,
	title = {Distributed and {Interactive} {Simulations} {Operating} at {Large} {Scale} for {Transcontinental} {Experimentation}},
	booktitle = {Distributed {Simulation} and {Real} {Time} {Applications} ({DS}-{RT}), 2010 {IEEE}/{ACM} 14th {International} {Symposium} on},
	author = {Gottschalk, Thomas D. and Yao, Ke-Thia and Wagenbreth, Gene and Lucas, Robert F. and Davis, Dan M.},
	month = oct,
	year = {2010},
	pages = {199--202},
}

@inproceedings{yao_data_2009,
	address = {Orlando, Florida},
	title = {Data {Analysis} for {Massively} {Distributed} {Simulations}},
	booktitle = {Interservice/{Industry} {Training}, {Simulation}, and {Education} {Conference} ({I}/{ITSEC}) 2009},
	author = {Yao, Ke-Thia and Lucas, Robert F. and Ward, Craig E. and Wagenbreth, Gene and Gottschalck, Thomas D.},
	year = {2009},
}

@inproceedings{tran_high_2008,
	address = {Orlando, Florida},
	title = {A {High} {Performance} {Route}-{Planning} {Technique} for {Dense} {Urban} simulations},
	booktitle = {Interservice/{Industry} {Training}, {Simulation}, and {Education} {Conference} ({I}/{ITSEC})},
	author = {Tran, John J. and Yao, Ke-Thia and Wagenbreth, Gene and Davis, Dan M. and Lucas, Robert F. and Bakeman, David J.},
	year = {2008},
}

@article{colon_modeling_2008,
	title = {Modeling {Human} {Perception} of {Situation} {Awareness} {During} {Constructive} {Experimentation}},
	volume = {3},
	number = {2},
	journal = {MSIAC Journal},
	author = {Colon, Philip and Tran, John and Yao, Ke-Thia and Anhalt, Michael and Curiel, Jacqueline},
	month = jul,
	year = {2008},
	pages = {20--29},
}

@inproceedings{colon_modeling_2007,
	address = {Orlando, Florida},
	title = {Modeling {Human} {Perception} of {Situation} {Awareness} {During} {Constructive} {Experimentation}},
	booktitle = {Interservice/{Industry} {Training}, {Simulation}, and {Education} {Conference} ({I}/{ITSEC})},
	author = {Colon, Philip and Anhalt, Michael and Curiel, Jacqueline and Yao, Ke-Thia and Tran, John},
	year = {2007},
}

@inproceedings{yao_agile_2006,
	address = {Orlando, Florida},
	title = {Agile {Data} {Logging} and {Analysis} for {Large}-{Scale} {Simulations}},
	booktitle = {Interservice/{Industry} {Training}, {Simulation}, and {Education} {Conference} ({I}/{ITSEC}) 2006},
	author = {Yao, Ke-Thia and Wagenbreth, Gene and Ward, Craig},
	year = {2006},
}

@article{yao_supercomputings_2006,
	title = {Supercomputing's {Role} in {Data} {Problems} and {Its} {Contribution} to {Solutions}},
	volume = {27},
	number = {3},
	journal = {The ITEA Journal of Test and Evaluation},
	author = {Yao, Ke-Thia and Davis, Dan and Lucas, Robert and Amburn, Philip},
	year = {2006},
	pages = {53--59},
}

@article{khoshnevis_mega-scale_2006,
	title = {Mega-scale fabrication by contour crafting},
	volume = {1},
	number = {3},
	journal = {International Journal of Industrial and Systems Engineering},
	author = {Khoshnevis, Behrokh and Hwang, Dooil and Yao, Ke-Thia and Yeh, Zhenghao},
	year = {2006},
}

@article{graebener_road_2005,
	title = {The {Road} to {Successful} {Joint} {Experimentation} {Starts} at the {Data} {Collection} {Trail} - {Part} {II}},
	journal = {MSIAC's M\&S Journal Online},
	author = {Graebener, Robert J. and Rafuse, Gregory and Miller, Robert and Yao, Ke-Thia},
	year = {2005},
}

@inproceedings{curiel_developing_2005,
	address = {Orlando, Florida},
	title = {Developing {Situation} {Awareness} {Metrics} in a {Synthetic} {Battlespace} {Environment}},
	booktitle = {Interservice/{Industry} {Training}, {Simulation}, and {Education} {Conference} ({I}/{ITSEC}) 2005},
	author = {Curiel, Jacqueline M. and Anhalt, Michael D. and Tran, John J. and Yao, Ke-Thia},
	year = {2005},
	file = {Curiel et al. - 2005 - Developing Situation Awareness Metrics in a Synthe.pdf:C\:\\Users\\ktyao\\Zotero\\storage\\H9MF4DRQ\\Curiel et al. - 2005 - Developing Situation Awareness Metrics in a Synthe.pdf:application/pdf},
}

@inproceedings{yao_simulation_2005,
	address = {Orlando, Florida},
	title = {Simulation {Data} {Grid}: {Joint} {Experimentation} {Data} {Management} and {Analysis}},
	booktitle = {Interservice/{Industry} {Training}, {Simulation}, and {Education} {Conference} ({I}/{ITSEC}) 2005},
	author = {Yao, Ke-Thia and Wagenbreth, Gene},
	year = {2005},
}

@inproceedings{wagenbreth_enabling_2005,
	title = {Enabling 1,000,000-{Entity} {Simulation} on {Distributed} {Linux} {Clusters}},
	booktitle = {Proceedings of the 2005 {Winter} {Simulation} {Conference}},
	author = {Wagenbreth, Gene and Yao, Ke-Thia and Davis, Dan M. and Lucas, Robert F. and Gottschalk, Thomas D.},
	editor = {Kuhl, M. E. and Steiger, N. M. and Armstrong, F. B. and Joines, J. A.},
	year = {2005},
}

@inproceedings{tran_interdisciplinary_2004,
	address = {Orlando, Florida},
	title = {An interdisciplinary approach to the study of battlefield simulation systems},
	booktitle = {Interservice/{Industry} {Training}, {Simulation}, and {Education} {Conference} ({I}/{ITSEC}) 2004},
	author = {Tran, John J. and Curiel, Jacqueline M. and Yao, Ke-Thia},
	month = dec,
	year = {2004},
}

@inproceedings{graebener_road_2004,
	address = {Orlando, Florida},
	title = {The {Road} to {Successful} {Joint} {Experimentation} {Starts} at the {Data} {Collection} {Trail} - {Part} {II}},
	booktitle = {Interservice/{Industry} {Training}, {Simulation}, and {Education} {Conference} ({I}/{ITSEC}) 2004},
	author = {Graebener, Robert J. and Rafuse, Gregory and Miller, Robert and Yao, Ke-Thia},
	month = dec,
	year = {2004},
}

@inproceedings{graebener_road_2003,
	address = {Orlando, Florida},
	title = {The {Road} to {Successful} {Joint} {Experimentation} {Starts} at the {Data} {Collection} {Trail}},
	booktitle = {Interservice/{Industry} {Training}, {Simulation}, and {Education} {Conference} ({I}/{ITSEC}) 2003},
	author = {Graebener, Robert J. and Rafuse, Gregory and Miller, Robert and Yao, Ke-Thia},
	month = dec,
	year = {2003},
}

@inproceedings{ko_dynamic_2002,
	address = {Honolulu Hawaii},
	title = {Dynamic coordination of information management services for processing dynamic web content},
	booktitle = {Proceedings of the {Eleventh} {International} {World} {Wide} {Web} {Conference}, {WWW2002}},
	publisher = {ACM},
	author = {Ko, In-Young and Yao, Ke-Thia and Neches, Robert},
	month = may,
	year = {2002},
	pages = {355--365},
}

@article{neches_geoworlds:_2001,
	title = {{GeoWorlds}: {Integrating} {GIS} and {Digital} {Libraries} for {Situation} {Understanding} and {Management}},
	volume = {7},
	journal = {The New Review of Hypermedia and Multimedia (NRHM)},
	author = {Neches, Robert and Yao, Ke-Thia and Ko, In-Young and Bugacov, Alejandro and Kumar, Vished and Eleish, Ragy},
	year = {2001},
	pages = {127--152},
}

@article{ko_semantic_2001,
	title = {A {Semantic} {Model} and {Composition} {Mechanism} for {Active} {Document} {Collection} {Templates} in {Web}-based {Information} {Management} {Systems}},
	volume = {5},
	number = {D},
	journal = {Electronic Transactions on Artificial Intelligence (ETAI)},
	author = {Ko, In-Young and Neches, Robert and Yao, Ke-Thia},
	year = {2001},
	pages = {55--71},
}

@inproceedings{yao_semantic_2001,
	address = {Brisbane, Australia},
	title = {Semantic {Interoperability} {Scripting} and {Measurements}},
	booktitle = {Working {Conference} on {Complex} and {Dynamic} {Systems} {Architecture}},
	author = {Yao, Ke-Thia and Ko, In-Young and Neches, Robert and MacGregor, Robert},
	month = dec,
	year = {2001},
}

@inproceedings{yao_asynchronous_2000,
	address = {San Antonio, TX},
	title = {Asynchronous information space analysis architecture using content and structure-based service brokering},
	booktitle = {Proceedings of the {Fifth} {ACM} {Conference} on {Digital} {Libraries}},
	author = {Yao, Ke-Thia and Ko, In-Young and Eleish, Ragy and Neches, Robert},
	year = {2000},
	pages = {133--142},
}

@inproceedings{ko_semantically-based_2000,
	address = {Lisbon, Portugal},
	title = {Semantically-{Based} {Active} {Document} {Collection} {Templates} for {Web} {Information} {Management} {Systems}},
	booktitle = {Proceedings of the {International} {Workshop} on the {Semantic} {Web}},
	author = {Ko, In-Young and Neches, Robert and Yao, Ke-Thia},
	month = sep,
	year = {2000},
}

@inproceedings{coutinho_geoworlds:_1999,
	address = {Orlando, Florida},
	title = {{GeoWorlds}: {A} {Geographically} {Based} {Situation} {Understanding} and {Information} {Management} {System} for {Disaster} {Relief} {Operations}},
	booktitle = {5th {International} {Conference} on {Information} {Systems}, {Analysis} and {Synthesis} ({ISAS} 99)},
	author = {Coutinho, Murilo and Neches, Robert and Bugacov, Alejandro and Kumar, Vished and Yao, Ke-Thia and Ko, In-Young and Eleish, Ragy and Abhinkar, Sameer},
	month = jul,
	year = {1999},
}

@inproceedings{coutinho_geoworlds:_1999-1,
	address = {Lyon, France},
	title = {{GeoWorlds}: {A} {Geographically} {Based} {Information} {System} for {Situation} {Understanding} and {Management}},
	booktitle = {First {International} {Workshop} on {TeleGeoProcessing} ({TeleGeo} 99)},
	author = {Coutinho, Murilo and Neches, Robert and Bugacov, Alejandro and Kumar, Vished and Yao, Ke-Thia and Ko, In-Young and Eleish, Ragy and Abhinkar, Sameer},
	month = may,
	year = {1999},
}

@inproceedings{yao_synchronous_1999,
	address = {University of Aizu, Fukushima, Japan},
	title = {Synchronous and {Asynchronous} {Collaborative} {Information} {Space} {Analysis} {Tools}},
	booktitle = {Proceedings of the {International} {Workshop} on {Collaboration} and {Mobile} {Computing} ({CMC} 99)},
	author = {Yao, Ke-Thia and Neches, Robert and Ko, In-Young and Eleish, Ragy and Abhinkar, {and} Sameer},
	month = sep,
	year = {1999},
}

@article{neches_collaborative_1998,
	title = {Collaborative {Information} {Space} {Analysis} {Tools}},
	journal = {D-Lib Magazine},
	author = {Neches, Robert and Abhinkar, Sameer and Hu, Fangqi and Eleish, Ragy and Ko, In-Young and Yao, Ke-Thia and Zhu, Quan and Will, Peter},
	month = oct,
	year = {1998},
}

@inproceedings{zhu_searching_1997,
	address = {Tsukuba Science City, Japan},
	title = {Searching for {Parts} and {Services} on the {Web}},
	url = {http://www.isi.edu/geoworlds/public/japan%20isdl%2097.pdf},
	booktitle = {Proceedings of the {International} {Symposium} on {Research}, {Development} and {Practice} in {Digital} {Libraries} ({ISDL} '97)},
	author = {Zhu, Quan and Hu, Fangqi and Yao, Ke-Thia and Will, Peter},
	month = nov,
	year = {1997},
}

@article{ellman_multi-level_1997,
	title = {Multi-level {Modeling} for {Engineering} {Design} {Optimization}},
	volume = {11},
	number = {5},
	journal = {AI EDAM},
	author = {Ellman, Tom and Keane, John and Schwabacher, Mark and Yao, Ke-Thia},
	year = {1997},
}

@article{yao_intelligent_1996,
	title = {Intelligent {Automated} {Grid} {Generation} {For} {Numerical} {Simulation}},
	volume = {10},
	number = {3},
	journal = {AI EDAM},
	author = {Yao, Ke-Thia and Gelsey, Andrew},
	year = {1996},
}

@inproceedings{yao_intelligent_1994,
	title = {Intelligent {Automated} {Grid} {Generation} for {Numerical} {Simulations}},
	booktitle = {{AAAI}},
	author = {Yao, Ke-Thia and Gelsey, Andrew},
	year = {1994},
	pages = {1224--1230},
}