{"id":379,"date":"2026-02-28T00:40:02","date_gmt":"2026-02-28T00:40:02","guid":{"rendered":"https:\/\/sites.wp.odu.edu\/huongq\/?page_id=379"},"modified":"2026-02-28T00:40:02","modified_gmt":"2026-02-28T00:40:02","slug":"national-science-foundation-internship","status":"publish","type":"page","link":"https:\/\/sites.wp.odu.edu\/huongq\/national-science-foundation-internship\/","title":{"rendered":"National Science Foundation internship"},"content":{"rendered":"\n
\n
\n

National Science Foundation
June 2024 – August 2024<\/p>\n<\/div>\n\n\n\n

\n
\"\"<\/figure>\n<\/div>\n<\/div>\n\n\n\n

Context<\/h2>\n\n\n\n

During my internship with the National Science Foundation (NSF) Research Experience, I worked on a project focused on developing adversarial patches capable of deceiving YOLO-based object detection systems. The project examined how modern surveillance and detection models can be manipulated through carefully designed visual perturbations. Over the course of the internship, I implemented and modified adversarial patch generation frameworks, adjusted loss functions, tested different initialization strategies, and evaluated performance across large datasets. This experience was significant because it introduced me to hands-on adversarial machine learning research and deepened my understanding of how AI systems can be vulnerable in real-world security applications.<\/p>\n\n\n\n

Artifact<\/h2>\n\n\n31-Shaped-Adversarial-Patches<\/a>\n

<\/p>\n\n\n\n

Reflection<\/h2>\n\n\n\n

In this role, I was responsible for implementing adversarial patch generation techniques and systematically testing their effectiveness against object detection models. I conducted experiments by adjusting optimization parameters, refining loss functions, and evaluating detection evasion performance on a large dataset. One of the main challenges I faced was understanding how theoretical adversarial attack concepts translated into practical implementation within the training pipeline. I addressed this by iteratively debugging the code, studying related research literature, and running controlled experiments to measure detection confidence and evasion rates.<\/p>\n\n\n\n

Through this internship, I learned how vulnerable AI systems can be when exposed to carefully crafted inputs. I strengthened my technical proficiency in PyTorch, machine learning model evaluation, and structured experimentation. More importantly, this experience shaped my professional identity by reinforcing my interest in adversarial machine learning and AI security. It taught me to think critically about system robustness rather than simply focusing on model performance.<\/p>\n\n\n\n

This experience connects directly to my broader professional development in cybersecurity and secure systems design. It strengthened my ability to analyze system weaknesses, support findings with data, and approach problems from both a technical and security-focused perspective. My time in the NSF research program represents a major step in developing from a student learning cybersecurity concept to a researcher actively testing and exposing system vulnerabilities.<\/p>\n\n\n\n

\n
\n

<\/p>\n<\/div>\n<\/div>\n\n\n\n

<\/p>\n","protected":false},"excerpt":{"rendered":"

National Science FoundationJune 2024 – August 2024 Context During my internship with the National Science Foundation (NSF) Research Experience, I worked on a project focused on developing adversarial patches capable of deceiving YOLO-based object detection systems. The project examined how modern surveillance and detection models can be manipulated through carefully designed visual perturbations. Over the… <\/p>\n

Read More<\/a><\/div>\n","protected":false},"author":30961,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"_links":{"self":[{"href":"https:\/\/sites.wp.odu.edu\/huongq\/wp-json\/wp\/v2\/pages\/379"}],"collection":[{"href":"https:\/\/sites.wp.odu.edu\/huongq\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/sites.wp.odu.edu\/huongq\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/sites.wp.odu.edu\/huongq\/wp-json\/wp\/v2\/users\/30961"}],"replies":[{"embeddable":true,"href":"https:\/\/sites.wp.odu.edu\/huongq\/wp-json\/wp\/v2\/comments?post=379"}],"version-history":[{"count":1,"href":"https:\/\/sites.wp.odu.edu\/huongq\/wp-json\/wp\/v2\/pages\/379\/revisions"}],"predecessor-version":[{"id":380,"href":"https:\/\/sites.wp.odu.edu\/huongq\/wp-json\/wp\/v2\/pages\/379\/revisions\/380"}],"wp:attachment":[{"href":"https:\/\/sites.wp.odu.edu\/huongq\/wp-json\/wp\/v2\/media?parent=379"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}