NSF Required Reporting

This material is based upon work supported by the National Science Foundation under Grant No. 1533768. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

For information about the duration and award amount, please refer to the award page.

Research challenges and project goals. This project concerns efficient parallel algorithms and software for emerging and future data analysis and mining applications, based on an emerging class of techniques known as tensor networks. Tensors, which are higher-dimensional generalizations of matrices, are finding applications in signal and image processing, computer vision, healthcare analytics, and neuroscience, to name just a few. Yet despite this demand, there is no comprehensive, high-performance software infrastructure targeting server systems that may have many parallel processors. Thus, the overarching research goal of this project is to design the first such infrastructure. The resulting prototype will be an open-source package, called the Parallel Tensor Infrastructure, or ParTI! The broader impact of the ParTI! project is to make the use of tensors, in a variety of data processing domains, much easier to do and more widespread.

Publications.

• Perros, Ioakeim; Chen, Robert; Vuduc, Richard; Sun, Jimeng. "Sparse Hierarchical Tucker Factorization and Its Application to Healthcare," Proceedings of the 2015 IEEE International Conference on Data Mining (ICDM), 2015. doi:10.1109/ICDM.2015.29
• Li, Jiajia; Battaglino, Casey; Perros, Ioakeim; Sun, Jimeng; Vuduc, Richard. "An input-adaptive and in-place approach to dense tensor-times-matrix multiply," Proceedings of SC'15: The ACM/IEEE International Conference on High-Performance Computing, Network, Storage, and Analysis, 2015. doi:10.1145/2807591.2807671
• Ioakeim Perros, Evangelos E. Papalexakis, Fei Wang, Richard W. Vuduc, Elizabeth Searles, Michael Thompson, Jimeng Sun. "SPARTan: Scalable PARAFAC2 for Large & Sparse Data," KDD, 2017, p. 375. doi:10.1145/3097983.3098014
• Ioakeim Perros, Fei Wang, Ping Zhang, Peter Walker, Richard W. Vuduc, Jyotishman Pathak, Jimeng Sun. "Polyadic Regression and its Application to Chemogenomics," SDM, 2017, p. 72. NSF-PAR
• Jiajia Li, Jee Choi, Ioakeim Perros, Jimeng Sun, Richard W. Vuduc. "Model-Driven Sparse CP Decomposition for Higher-Order Tensors," IPDPS, 2017, p. 1048. doi:10.1109/IPDPS.2017.80
• Ardavan Afshar and Joyce Ho and Bistra Dilkina and Ioakeim Perros and Elias Khalil and Li Xiong and Vaidy Sunderam. "CP-ORTHO: An orthogonal tensor factorization framework for spatio-temporal data," Proceedings of the International Conference on Advances in Geographic Information Systems (SIGSPATIAL), 2017. doi:10.1145/3139958.3140047
• Ioakeim Perros, Evangelos E. Papalexakis, Haesun Park, Richard Vuduc, Xiaowei Yan, Christopher deFilippi, Walter F. Stewart, Jimeng Sun (2018). SUSTain: Scalable Unsupervised Scoring for Tensors and its Application to Phenotyping. Proceedings of the SIGKDD Conference on Knowledge Discovery and Data Mining (KDD). doi:10.1145/3219819.3219999
• Jiajia Li, Jimeng Sun, Richard Vuduc. HiCOO: Hierarchical storage of sparse tensors. Proceedings of the ACM/IEEE International Conference for High-Performance Computing, Networking, Storage, and Analysis (``Supercomputing'' or SC), November 2018. doi:10.1109/SC.2018.00022. Winner, Best Student Paper
• Nisa, Israt; Li, Jiajia; Sukumaran-Rajam, Aravind; Vuduc, Richard; Sadayappan, P. Load-balanced sparse MTTKRP on GPUs. IPDPS'19: Proceedings of the International Parallel and Distributed Processing Symposium. May 2019. doi:10.1109/IPDPS.2019.00023
• Li, Jiajia; Uçar, Bora; Çatalyürek, Ümit; Sun, Jimeng; Vuduc, Richard. Efficient and effective sparse tensor reordering. ICS '19: Proceedings of the ACM International Conference on Supercomputing (ICS), June 2019. doi:10.1145/3330345.3330366
• Yuchen Ma, Jiajia Li, Chenggang Yang, Jimeng Sun, Richard Vuduc. Optimizing sparse tensor times matrix on GPUs. J. Parallel and Distributed Computing (JPDC), 129:99--109, July 2019. doi:10.1016/j.jpdc.2018.07.018
• Perros, Ioakeim Perros; Papalexakis, Evangelos E.; Vuduc, Richard W.; Eliza Searles, Elizabeth; Sun, Jimeng. Temporal phenotyping of medically complex children via PARAFAC2 tensor factorization. Journal of Biomedical Informatics, 2019. doi:10.1016/j.jbi.2019.103125

Software products.

• ParTI! -- http://parti-project.org
• SUSTAIN -- https://github.com/kperros/SUSTain

Datasets.

• FROSTT -- http://frostt.io/

Dissertations.

• Ioakeim Perros. Knowledge Discovery from Multidimensional Health Data: Methods and Applications. Ph.D. Dissertation, Georgia Institute of Technology (2019).
• Jiajia Li. Scalable tensor decompositions in high-performance computing environments. Ph.D. Dissertation, Georgia Institute of Technology (2018).

Tutorials.

• Xiao, Cao; Sun, Jimeng (2019). Tutorial: Data Mining Methods for Drug Discovery and Development. KDD '19: Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. doi:10.1145/3292500.3332273