SmartTensors

Colorful banner graphic

SmartTensors

Unsupervised AI for Big-Data Analytics

SmartTensors is a groundbreaking, unsupervised Artificial Intelligence (AI) methodology and software suite for latent feature discovery and predictions in big data.

The only tensor codes to:

  • identify and extract latent features in very large data sets
  • offer explainable machine learning
  • make informative, robust predictions
  • determine dependencies automatically

GitHub Codes

pyDNMFk | DnMFKpyOBTNs | pyCP-APR | HNMF | pyDNTNK


Applications
  • Medicine: latent patterns in cancer genomics, metabolomics, protein structure.
  • Economy: macro-economy analyses, marketing.
  • Chemistry: discovering new chemical pathways and reactions, phase separation analysis in complex liquids, co-polymers, cell-membranes.
  • Climate: ice and water masses transient patterns, micro-climate patterns, land ice pattern.
  • Material Science: analysis of combinatorial material libraries based on their: X-ray, Raman fluorescence and other spectra.
  • Text Mining: topic modeling, topic evolution.
  • Agriculture: estimating the role of different artificial substances on the yield.
  • Computer security: anomaly and fraud detection.
  • Blind source separation: detection of anomaly-diffusion and particles distribution from dynamic light scattering.
  • Relational databases: Boolean factorization analysis of categorical patterns.
  • Dynamic Networks and Ranking: detection of latent communities in directed and undirected graphs and networks, ranking of latent research communities hidden in temporal multilayer networks.
  • Data Compression: compression of large images and videos (e.g. asteroid water impacts), scientific computer-generated data, etc.
  • Subsurface: contamination, hidden sources.
  • Seismology: artificial earthquakes, industrial noise.

Team

Publications
References in support of the R&D100 application
  1. "COVID-19 multidimensional Kaggle literature organization." Eren, Maksim E., Nick Solovyev, Chris Hamer, Renee McDonald, Boian S. Alexandrov, and Charles Nicholas. In Proceedings of the 21st ACM Symposium on Document Engineering, pp. 1-4. 2021.
  2. Quantum Annealing Algorithms for Boolean Tensor Networks. Pelofske E, Hahn G, O'Malley D, Djidjev HN, Alexandrov BS. arXiv preprint arXiv:2107.13659. 2021 Jul 28.
  3. Selection of Optimal Salient Time Steps by Non-negative Tucker Tensor Decomposition, Jesus Pulido, John Patchett, Manish Bhattarai, Boian Alexandrov, and James Ahrens,  in EuroVis 2021. 2021. Hosted by, University of Zurich in collaboration with FAU Erlangen-Nuremberg and ETH Zurich:
  4. Boolean Hierarchical Tucker Networks on Quantum Annealers
  5. Elijah Pelofske, Georg Hahn, Daniel O'Malley, Hristo N. Djidjev, Boian S. Alexandrov
  6. Source identification by non-negative matrix factorization combined with semi-supervised clustering, BS. Alexandrov, LB. Alexandrov et al. US Patent S10,776,718, Sep. 2020
  7. Distributed Non-Negative Tensor Train Decomposition, M Bhattarai, G Chennupati, E Skau, R Vangara, H Djidjev, BS Alexandrov, to appear in proceedings of HPEC:  2020 IEEE High Performance Extreme Computing Virtual Conference September 2020
  8. Tucker-1 Boolean Tensor Factorization with Quantum Annealers, Daniel O'Malley, Hristo Djidjev and Boian Alexandrov, to appear in proceedings of IEEE International Conference on Rebooting Computing (ICRC 2020), 1-3 December, 2020
  9. Semantic Nonnegative Matrix Factorization with Automatic Model Determination for Topic Modeling, R Vangara, E. Skau, G Chennupati, et al., proceedings of 19th IEEE International Conference on Machine Learning and Applications, December 14-17, 2020
  10. Multi-Dimensional Anomalous Entity Detection via Poisson Tensor Factorization, Maksim Eren, Juston Moore, Boian S. Alexandrov, to appear in proceedings of 18th IEEE International Conference on Intelligence and Security Informatics (ISI), Nov. 9-10, 2020 
  11. A neural network for determination of latent dimensionality in Nonnegative Matrix Factorization, B Nebgen, R Vangara, MA Hombrados-Herrera, S Kuksova, BS Alexandrov, Journal of Machine Learning: Science and Technology, 2020.
  12. An out of memory tSVD for big-data factorization, H Carrillo-Cabada, E Skau, G Chennupati, BS Alexandrov, H Djidjev, IEEE Access 8, 107749-107759, 2020.
  13. Identification of anomalous diffusion sources by unsupervised learning, R Vangara, KØ Rasmussen, DN Petsev, G Bel, BS Alexandrov, Physical Review Research 2 (2), 023248
  14. Weak matching of temporal interval graphs of sensors for robust multi-modal event detection in noise, L Prasad, BS Alexandrov, BT Nebgen in proceedings of Signal Processing, Sensor/Information Fusion, and Target Recognition XXIX, 2020.
  15. Determination of Latent Dimensionality in International Trade Flow, DP Truong, E Skau, VI Valtchinov, BS Alexandrov, Mach. Learn.: Sci. Technol. 1 045017, 2020.
  16. Metabolomics and the pig model reveal aberrant cardiac energy metabolism in metabolic syndrome, M Karimi, V Petkova, JM Asara, MJ Griffin, FW Sellke, et, all., Scientific reports 10 (1), 1-11, 2020.
  17. Distributed non-negative matrix factorization with determination of the number of latent features, G Chennupati, R Vangara, E Skau, H Djidjev, B Alexandrov, The Journal of Supercomputing, 1-31, 2020.
  18. Coarse-Grain Cluster Analysis of Tensors With Application to Climate Biome Identification, D DeSantis, PJ Wolfram, K Bennett, B Alexandrov, Journal of Machine Learning: Science and Technology, 2020.
  19. The genomic and epigenomic evolutionary history of papillary renal cell carcinomas, M Zhu, B., Poeta, M.L., Costantini, e al., Nature Communication, 3096, 2020
  20. Increased irrigation water salinity enhances nitrate transport to deep unsaturated soil, G Weissman, G Bel, A Ben‐Gal, U Yermiyahu, B Alexandrov, et al., Vadose Zone Journal 19 (1), e2004, 2020
  21. Robust effect of metabolic syndrome on major metabolic pathways in the myocardium, M Karimi, VI Pavlov, O Ziegler, N Sriram, SY Yoon, V Agbortoko, et al., Plos one 14 (12), e0225857, 2019.
  22. Non-Negative Matrix Factorization for Selection of Near-Native Protein Tertiary Structures, N Akhter, R Vangara, G Chennupati, BS Alexandrov, H Djidjev, A Shehu, in the proceedings of 2019 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)
  23. Unsupervised machine learning based on non-negative tensor factorization for analyzing reactive-mixing, VV Vesselinov, MK Mudunuru, S Karra, D O'Malley, BS Alexandrov, Journal of Computational Physics 395, 85-104, 2019.
  24. Nonnegative Canonical Polyadic Decomposition with Rank Deficient Factors, B Alexandrov, D DeSantis, G Manzini, E Skau, arXiv preprint arXiv:1909.07570, 2019.
  25. Unsupervised Machine Learning for Analysis of Phase Separation in Ternary Lipid Mixture, CA Löpez, VV Vesselinov, S Gnanakaran, BS Alexandrov, Journal of Chemical Theory and Computation 15 (11), 6343-6357
  26. Nonnegative tensor decomposition with custom clustering for microphase separation of block copolymers, BS Alexandrov, VG Stanev, VV Vesselinov, KØ Rasmussen, Statistical Analysis and Data Mining: The ASA Data Science Journal 12, 2019.
  27. Nonnegative tensor factorization for contaminant source identification, VV Vesselinov, BS Alexandrov, D O'Malley, Journal of contaminant hydrology 220, 66-97
  28. Nonnegative/binary matrix factorization with a d-wave quantum annealer, D O’Malley, VV Vesselinov, BS Alexandrov, LB Alexandrov, PloS one 13 (12), e0206653, 2018.
  29. Unsupervised Phase Mapping of X-ray Diffraction Data by Nonnegative Matrix, Factorization Integrated with Custom Clustering, BSA Valentin Stanev, Velimir V. Vesselinov, A. Gilad Kusne, Graham Antoszewski, npj Computational Materials 4 (43), 2018.
  30. Contaminant source identification using semi-supervised machine learning, VV Vesselinov, BS Alexandrov, D O’Malley, Journal of contaminant hydrology 212, 134-142
  31. Nonnegative Matrix Factorization for identification of unknown number of sources emitting delayed signals, FL Iliev, VG Stanev, VV Vesselinov, BS Alexandrov, PloS one 13 (3), e0193974, 2018.
  32. Identification of release sources in advection-diffusion system by machine learning combined with Green’s function inverse method, V Stanev, F Iliev, S Hansen, V Velimir, B Alexandrov, in Applied Mathematical Modelling, Volume 60, August 2018, Pages 64-76