Akash Vartak

Degree: PhD, Computer Science

Looking for: Internships

I am pursuing my PhD in Computer Science @ University Of Maryland - Baltimore County in the Dept. of Computer Science and Electrical Engineering.

I work as a Research Assistant to Prof. Tim Oates in the CORAL Lab and work with Graph Neural Networks for deep learning model classification and adversarial model detection. My past research includes using traditional CNN architectures and NLP techniques for Grounding Language using Visual Percepts.

I am open to internship opportunities as a ML/AI Intern or in a research position. I am also open to collaborate on research projects or paper ideas. You can reach me at the email. View CV

news

01 Jun 2026 Interning at Synaptiq as a Data Science and AI Engineering Intern.
06 Feb 2026 Released the Final Report of the IARPA + NIST program Trojans in Artificial Intelligence (TrojAI), a multi-year effort to detect hidden backdoors in AI models.
15 Nov 2025 Paper on using an adaptive LLM-driven label-free for open-vocabulary object detection presented at IEEE Baltimore Technical and Professional Development Colloquium 2025
02 Jul 2025 Paper on using GCNs as meta-classifiers to detect trojaned DNN models accepted at International Joint Conference on Neural Networks 2025
26 Feb 2025 Paper on using GCNs as meta-classifiers to detect trojaned DNN models published on arXiv
06 Feb 2025 Successfully passed Comprehensive Portfolio - 1^st milestone towards PhD.
31 Aug 2022 Started PhD in Computer Science at UMBC.
24 Mar 2022 Successfully defended Masters Thesis.
28 Aug 2019 Started Masters in Computer Science at UMBC.

selected publications

arXiv

Trojans in Artificial Intelligence (TrojAI) Final Report

Kristopher W. Reese, Taylor Kulp-McDowall, Michael Majurski, Tim Blattner, Derek Juba, Peter Bajcsy, Antonio Cardone, Philippe Dessauw, Alden Dima, Anthony J. Kearsley, Melinda Kleczynski, Joel Vasanth, Walid Keyrouz, Chace Ashcraft, Neil Fendley, Ted Staley, Trevor Stout, Josh Carney, Greg Canal, Will Redman, Aurora Schmidt, Cameron Hickert, William Paul, Jared Markowitz, Nathan Drenkow, David Shriver, Marissa Connor, Keltin Grimes, Marco Christiani, Hayden Moore, Jordan Widjaja, Kasimir Gabert, Uma Balakrishnan, Satyanadh Gundimada, John Jacobellis, Sandya Lakkur, Vitus Leung, Jon Roose, Casey Battaglino, Farinaz Koushanfar, Greg Fields, Xihe Gu, Yaman Jandali, Xinqiao Zhang, Tara Javidi, Akash Vartak, Tim Oates, Ben Erichson, Michael Mahoney, Rauf Izmailov, Xiangyu Zhang, Guangyu Shen, Siyuan Cheng, Shiqing Ma, XiaoFeng Wang, Haixu Tang, Di Tang, Xiaoyi Chen, Zihao Wang, Rui Zhu, Susmit Jha, Xiao Lin, Manoj Acharya, Weichao Zhou, Feisi Fu, Panagiota Kiourti, Chenyu Wang, Zijian Guo, H M Sabbir Ahmad, Wenchao Li, and Chao Chen

2026

Bib HTML PDF

@misc{reese2026trojansartificialintelligencetrojai,
  title = {Trojans in Artificial Intelligence (TrojAI) Final Report},
  author = {Reese, Kristopher W. and Kulp-McDowall, Taylor and Majurski, Michael and Blattner, Tim and Juba, Derek and Bajcsy, Peter and Cardone, Antonio and Dessauw, Philippe and Dima, Alden and Kearsley, Anthony J. and Kleczynski, Melinda and Vasanth, Joel and Keyrouz, Walid and Ashcraft, Chace and Fendley, Neil and Staley, Ted and Stout, Trevor and Carney, Josh and Canal, Greg and Redman, Will and Schmidt, Aurora and Hickert, Cameron and Paul, William and Markowitz, Jared and Drenkow, Nathan and Shriver, David and Connor, Marissa and Grimes, Keltin and Christiani, Marco and Moore, Hayden and Widjaja, Jordan and Gabert, Kasimir and Balakrishnan, Uma and Gundimada, Satyanadh and Jacobellis, John and Lakkur, Sandya and Leung, Vitus and Roose, Jon and Battaglino, Casey and Koushanfar, Farinaz and Fields, Greg and Gu, Xihe and Jandali, Yaman and Zhang, Xinqiao and Javidi, Tara and Vartak, Akash and Oates, Tim and Erichson, Ben and Mahoney, Michael and Izmailov, Rauf and Zhang, Xiangyu and Shen, Guangyu and Cheng, Siyuan and Ma, Shiqing and Wang, XiaoFeng and Tang, Haixu and Tang, Di and Chen, Xiaoyi and Wang, Zihao and Zhu, Rui and Jha, Susmit and Lin, Xiao and Acharya, Manoj and Zhou, Weichao and Fu, Feisi and Kiourti, Panagiota and Wang, Chenyu and Guo, Zijian and Ahmad, H M Sabbir and Li, Wenchao and Chen, Chao},
  year = {2026},
  eprint = {2602.07152},
  archiveprefix = {arXiv},
  primaryclass = {cs.CR},
  url = {https://arxiv.org/abs/2602.07152},
  language = {English},
}

IJCNN
DeBUGCN – Detecting Backdoors in CNNs Using Graph Convolutional Networks

Akash Vartak, Khondoker Murad Hossain, and Tim Oates

In , 2025

Abs Bib HTML PDF

Deep neural networks (DNNs) are becoming commonplace in critical applications, making their susceptibility to backdoor (trojan) attacks a significant problem. In this paper, we introduce a novel backdoor attack detection pipeline, detecting attacked models using graph convolution networks (DeBUGCN). To the best of our knowledge, ours is the first use of GCNs for trojan detection. We use the static weights of a DNN to create a graph structure of its layers. A GCN is then used as a binary classifier on these graphs, yielding a trojan or clean determination for the DNN. To demonstrate the efficacy of our pipeline, we train hundreds of clean and trojaned CNN models on the MNIST handwritten digits and CIFAR-10 image datasets, and show the DNN classification results using DeBUGCN. For a true In-the-Wild use case, our pipeline is evaluated on the TrojAI dataset which consists of various CNN architectures, thus showing the robustness and model-agnostic behaviour of DeBUGCN. Furthermore, on comparing our results on several datasets with state-of-the-art trojan detection algorithms, DeBUGCN is faster and more accurate.
@inproceedings{vartak2025debugijcnn, title = {DeBUGCN -- Detecting Backdoors in CNNs Using Graph Convolutional Networks}, author = {Vartak, Akash and Hossain, Khondoker Murad and Oates, Tim}, year = {2025}, conference = {International Joint Conference on Neural Networks}, url = {https://ieeexplore.ieee.org/abstract/document/11227981}, language = {English}, }
Masters Thesis
Using Text to Improve Classification of Man-Made Objects

Akash A. Vartak

2022

Copyright - Database copyright ProQuest LLC; ProQuest does not claim copyright in the individual underlying works; Last updated - 2023-08-04

Abs Bib HTML PDF

People identify man-made objects by their visual appearance and the text on them e.g., does a bottle say water or shampoo? We use text as an important visual cue to help distinguish between similar looking objects. This thesis explores a novel joint model of visual appearance and textual cues for image classification.We perform this in three functions - (a) Isolating an object in an input image; (b) Extracting text from the image; (c) Training a joint vision/text model. We simplify the task by extracting text separately and presenting it to the model in machine readable format. Such a joint model has utility in many real world challenges where language is interpreted through a sensory perception like vision or sound.The aim of the research is to understand whether visual percepts, when understood in the context of extracted language, will provide a better classification of image objects than using only pure vision to perform image classification. In conclusion, we show that joint classifier models can successfully make use of text present in images to classify objects, provided that the extracted text from images is of high quality and we have the number of images proportional to the number of classification classes.
@phdthesis{vartak2022thesis, author = {Vartak, Akash A.}, year = {2022}, title = {Using Text to Improve Classification of Man-Made Objects}, journal = {ProQuest Dissertations and Theses}, pages = {55}, note = {Copyright - Database copyright ProQuest LLC; ProQuest does not claim copyright in the individual underlying works; Last updated - 2023-08-04}, keywords = {Classification; Computer vision; Deep learning; Machine learning; Natural language processing; Text extraction; Computer science; 0984:Computer science}, isbn = {9798837525858}, language = {English}, url = {http://proxy-bc.researchport.umd.edu/login?url=https://www.proquest.com/dissertations-theses/using-text-improve-classification-man-made/docview/2695026845/se-2}, }