Shane Dirksen

About

Hello, I'm Shane Dirksen, I'm from California, I'm a US Army veteran, and I hold a Bachelor's Degree in Computer Science from Cal Poly Pomona. I have a tremendous interest in AI research, particularly with deep learning applications like computer vision.

Projects

Wildfire Spread Prediction

Given limited imagery from high altitude cameras and sensors, this project aims to both detect wildfires and predict fire spread.

In this video, the smoke detection model outputs a mask, dilates it, and uses sparse optical flow. Given the camera angle, the Ground Sampling Distance (GSD), and framerate, the system can determine wind speed and direction. This then is fed into the overal fire spread prediction model.

Python NVIDIA Omniverse

---

Code: Code is Currently Not Released to the Public

Anomaly Detection using Graph-Based Deep Learning

Using the latent representations and the reconstruction error as the graph vertices, with cosine similarity to represent the edges, the model can then identify which class samples belong to unknown classes. This research also introduces a novel loss, which aims to improve the modularity of the known classes within the graph.

Python

---

Code: Github

Action Recognition Networks and Human Saliency

Reseach project that focuses on comparing and analyzing what human observers and action recognition networks observe when tasked with identifying an action in a video.

Python

---

Code: Code is Currently Not Released to the Public

Semi-Synthetic Dataset Generation

This project generates large, unique image datasets using real world imagery for backgrounds and synthetic objects that are placed within the scene. The dataset also generates coordinates and labels for the respective bounding boxes of the objects.

Python MEL Maya 2022

---

Code: Github

Resume

Education

• B.S. in Computer Science. GPA: 3.91
• Undergraduate Coursework: Artificial Intelligence; Machine Learning; Systems Programming; Numerical Methods; Concepts of Programming Languages; Software Engineering; C++ Programming; Computer Architecture; Computer Networks.
• Scholarships: Cal-Bridge Scholarship; Sally Casanova Pre-Doctoral Scholarship; Last Mile Fellowship; Boots to Broncos Scholarship.

• AS-T in Computer Science, May 2021. GPA: 3.87
• Undergraduate Coursework: Intro to Algorithms and Programming/Java; Architecture/Assembly Language; Mechanics of Solids and Fluids; Data Structures/Programming Design; Physics: Electricity/Magnetism; Calculus II.
• Scholarship: Chief Information System Officers Association Scholarship.

Employment

• Conducted research on satellite-based wildfire detection and prediction using deep learning. Python
• Developed an optical flow algorithm to ascertain wind speed and direction; implemented a digital twin platform for dynamic wildfire growth influenced by wind vectors.

• Research in computer vision with action recognition networks. Python
• Used data from human observers when tasked with classifying a video and compared with an action recognition network to make analysis and comparisons.

• Responsible for conducting inspections on Security Officers at various sites.
• Site Supervisor for 3 college campuses, responsible for 18 Security Officers, and 6 Student Workers.

• Signed for over $400k worth of equipment as a machine gunner; team leader of four soldiers.
• Responded to 14 emergency response situations, tracked over 700 convoys while deployed in the Balkans.

Research Experience

• Anomaly Detection via STARS (Student Success and Transfer Articulation through Research and Support Services) (2022–Current): Developed a novel anomaly detection framework combining autoencoders and graph-based techniques. Implemented a model that encodes the latent space representations and reconstruction error into a graph to identify unknown classes. Python
• Semi-Synthetic Image Dataset Generation via RIO (Research through Inclusive Opportunities) Program (2021–2022): Conducted computer vision research. Generated a data set using 3D modeling software. A script generates unique, realistic pictures and a python program detects the bounding boxes. The data set is used to train a deep neural network model for object detection. Python, MEL
• Autonomous UAV Research (2021–2022): Autonomous Navigation of UAVs in Indoor Environments for Search and Rescue Missions (Use simultaneous localization and mapping (SLAM) and other techniques for autonomous navigation of UAVs in the GPS-denied indoor environments). Python
• Cryptocurrency Price Analysis (2021–2022): Conducting various analyses to identify price indicators related to cryptocurrencies, such as co-variance analysis and frequency analysis. Python

Publications

• E. Ngo, J. Ramirez, M. Medina-Soto, S. Dirksen, E. D. Victoriano and S. Bhandari, ”UAV Platforms for Autonomous Navigation in GPS-Denied Environments for Search and Rescue Missions,” 2022 International Conference on Unmanned Aircraft Systems (ICUAS), 2022, pp. 1481-1488, doi: 10.1109/ICUAS54217.2022.9836181.

Presentations

• Dirksen, S., Ernesto, C., & Korah, J. (Spring, 2023). Using Explainable AI with Deep Autoencoders and Recurrent Neural Networks for Cyber Anomaly Detection. Science Symposium.
• Dirksen, S., & Korah, J. (Fall, 2022). Removing the Black Box: Parallelization of Neural-Backed Decision Trees for Explanation Based Cyber Anomaly Detection. Southern California Conferences for Undergraduate Research.
• Morgan, M., Dirksen, S., Chang, R., & Wloka, C. (Summer, 2022). Do Machine Learning Methods for Action Recognition Focus on the Same Video Elements as Human Viewers? USC REU Symposium.
• Dirksen, S., & Ji, H. (Spring, 2022). Generation of Semi-Synthetic Image Datasets for Training Deep Learning Models. Research, Scholarship & Creative Activities Conference.
• Lee, J., Barrientos, J., Townsend, K., Dirksen, S., Ullum, K., & Eger, M. (Spring 2022). Carnival Games using Virtual Reality. Science Symposium.
• Dirksen, S., Reddy, V., & Korah, J. (Spring, 2022). Parallel Designs for Machine Learning Based Cybersecurity Anomaly Detection. Science Symposium.

Professional Development

Academic Affiliations

• Google Developers Club CPP: Vice President of Administration, interim President. Responsible for scheduling and managing all club events and meetings, as well as managing all club finances.
• Forensics and Security Technology Club: Member of the Blockchain Research Team. Responsible for cryptography research. Presented findings and provided instruction to club members.

Research Programs

• Cal-Bridge Scholar (2022–Current): Selected as a scholar for the Cal-Bridge program. Attended various workshops and mentorship programs.
• CS Research Mentorship Program (CSRMP), Google Research (2023). Google sponsored research mentorship program.
• Office of Undergrad Research (OUR) GRAD Program (2021–2022): OUR Grad Scholar.

Additional Education

• Data Science Immersion (DSI), Divergence Academy (2023). Data Science bootcamp; studied tools such as PowerBI, Databricks, and SQL.
• CodePath’s Cybersecurity Course (Spring 2022). Explored web application security, vulnerability exploitation, and simulated real-world cyber threats.
• Big Data and Cloud Computing Apprenticeship (2021–2022). Received training in utilizing various cloud computing platforms. Created a stock/crypto price prediction via machine learning and investment automation.
• Achieve Scholars (ASP) Program (2021–2022): ASP Scholar.

Languages and Software

• C, C++, C#, Java, MEL, SQL, Python, Ruby
• Visual Studio, MySQL, PyCharm, Unity, NetBeans, Maya 2022, NVIDIA Omniverse
• Amazon Web services, Google Cloud Platform, Microsoft Azure

Elements

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';