Biography
I am a graduate Computer Science student and a SFS CyberCorps alumni at the University of Maryland, Baltimore County (UMBC). I received a BS in Computer Science Summa Cum Laude at UMBC. I work as a Research Assitant at Los Alamos National Laboratory (LANL) and at UMBC’s DREAM Lab. My research interests lie at the intersection of the machine learning and cybersecurity disciplines, with a concentration in tensor decomposition.
Interests
- Machine Learning
- Tensor Decompositions
- Cybersecurity
Education
MS in Computer Science, 2022
University of Maryland, Baltimore County (UMBC)
BS in Computer Science, 2020
University of Maryland, Baltimore County (UMBC)
AA in Computer Science, 2018
Montgomery College (MC)
Recent Publications
(2021).
COVID-19 Multidimensional Kaggle Literature Organization.
In ACM Symposium on Document Engineering 2021 (DocEng ’21), 2021.
(2021).
Evading Malware Classifiers via Monte Carlo Mutant Feature Discovery.
Presented at the 12th Annual Malware Technical Exchange Meeting, Online, 2021.
(2021).
Random Forest of Tensors (RFoT).
Presented at the 12th Annual Malware Technical Exchange Meeting, Online, 2021.
(2020).
Multi-Dimensional Anomalous Entity Detection via Poisson Tensor Factorization.
In IEEE International Conference on Intelligence and Security Informatics (ISI), 2020.
(2020).
COVID-19 Kaggle Literature Organization.
In ACM Symposium on Document Engineering 2020 (DocEng ’20), 2020.
DOI Preprint URL Preprint PDF Poster Code Interactive Plot Details
Recent Talks
Anomalous Event Detection using Non-Negative Poisson Tensor Factorization
Network intrusion detection systems that are based on statistical User Behaviour Analytics play a fundamental role in the identification of anomalous agents such as malicious insiders, misused accounts, and users with compromised credentials. To this extent, there have been significant results in detecting anomalies from learned user behavior models via non-negative Poisson matrix factorization. We expand upon previous work in this project by exploiting the higher dimensional and sparse problems created by the user authentication data. An integrated multidimensional anomaly scoring method based on tensors and Poisson recommender systems is proposed. In our experiments, we build a higher-order model that can detect the accounts compromised by red-team during penetration testing activities at a large organization.
Software
lanl/pyDNTNK
pyDNTNK is a software package for applying non-negative Hierarchical Tensor decompositions such as Tensor train and Hierarchical Tucker decompositons in a distributed fashion to large datasets. It is built on top of pyDNMFk. Tensor train (TT) and Hierarchical Tucker(HT) are state-of-the-art tensor network introduced for factorization of high-dimensional tensors. These methods transform the initial high-dimensional tensor in a network of low dimensional tensors that requires only a linear storage. Many real-world data,such as, density, temperature, population, probability, etc., are non-negative and for an easy interpretation, the algorithms preserving non-negativity are preferred. Here, we introduce the distributed non-negative Hierarchical tensor decomposition tools and demonstrate their scalability and the compression on synthetic and real world big datasets. pyDNTNK is developed as part of the R&D 100 award wining SmartTensors project.
lanl/pyQBTNs
pyQBTNs is a Python library for boolean matrix and tensor factorization using D-Wave quantum annealers. The library includes five different boolean tensor decomposition methods making up three distinct types of tensor networks. pyQBTNs is developed as part of the R&D 100 award wining SmartTensors project.
lanl/pyCP_APR
pyCP_APR is a Python library for tensor decomposition and anomaly detection that is developed as part of the R&D 100 award wining SmartTensors project. It is designed for the fast analysis of large datasets by accelerating computation speed using GPUs. pyCP_APR uses the CANDECOMP/PARAFAC Alternating Poisson Regression (CP-APR) tensor factorization algorithm utilizing both Numpy and PyTorch backend. While the Numpy backend can be used for the analysis of both sparse and dense tensors, PyTorch backend provides faster decomposition of large and sparse tensors on the GPU. pyCP_APR’s Scikit-learn like API allows comfortable interaction with the library, and include the methods for anomaly detection via the p-values obtained from the CP-APR factorization.
lanl/pyDNMFk
pyDNMFk is a software package for applying non-negative matrix factorization in a distributed fashion to large datasets. It has the ability to minimize the difference between reconstructed data and the original data through various norms (Frobenious, KL-divergence). Additionally, the Custom Clustering algorithm allows for automated determination for the number of Latent features. pyDNMFk is developed as part of the R&D 100 award wining SmartTensors project.