NEWS
I have joined CoreLogic as a Senior Machine Learning Scientist, working on GenAI projects. (10/2024)
EDUCATION
- UC San Diego, PhD, Computer Engineering (advised by Prof. Farinaz Koushanfar) (Aug. 2024)
- San Diego State University, MSEE, Computer Engineering (Dec. 2019)
- Northeastern University (CN), BSEE, Computer Engineering (May 2017)
EXPERIENCE
Founder and Chief Technology Officer (CTO), Check-It Analytics, San Diego, CA (Aug. 2023 - Current)
- Founded and led the development of an AI-driven financial information platform to address the time-consuming issue of collecting and analyzing financial news.
- Used RAG for LLMs to streamline financial processes and provide customers with suggested questions and answers.
- Achieved up to 80% time savings compared to traditional financial platforms.
Research Intern, Arm, Austin, TX (June 2023 - Sep. 2023)
- Address the challenge of large data size without compromising efficiency.
- Developed a data distillation algorithm.
- Reduced data size to at least 1/10,000 and improved ML model performance by at least 50%.
Research Intern, Arm, Austin, TX (May 2022 - Aug. 2022)
- Led a SoC trace data distillation project to improve data processing efficiency.
- Developed a novel vulnerability detection algorithm using Graph Neural Networks (GNNs).
- Reduced data processing times by 50%.
Graduate Student Researcher, UC San Diego, La Jolla, CA (Dec. 2019 - Current)
- Developed innovative techniques for identifying compromised artificial intelligence models and enhancing security.
- Played a leading role in a team that achieved 2nd place among 16 competitors in a notable AI security challenge
Deacon Board Member, Fresh Wind Chinese Church of San Diego, San Diego, CA (Aug. 2019 - Current)
- Led the outreach department to organize large-scale social events
- Successfully organized events for up to 120 people
HIGHLIGHTED PUBLICATION & PATENTS
- (Trustworthy ML) Z. Ghodsi*, M, Javaheripi*, N. Sheyban*, X. Zhang*, K, Huang, & F. Koushanfar, (2023).zPROBE: Zero Peek Robustness Checks for Federated Learning, (ICCV23)
- [PATENT: zPROBE: Zero Peek Robustness Checks for Federated Learning] 2022 (Serial No.63/496,157.)
-
(Trustworthy ML) Z. Ghodsi*, M, Javaheripi*, N. Sheybani*,X. Zhang*, K, Huang, & F. Koushanfar, (2022).zPROBE: Zero Peek Robustness Checks for Federated Learning, (NeurIPS’22-TSRML) [Outstanding paper award]
-
(Trustworthy ML) X. Zhang, M. Samragh, S. Hussain, K. Huang, & F. Koushanfar. Scalable Binary Neural Network applications in Oblivious Inference, (ACM TECS)
-
(Watermarking) P. Neekhara, S. Hussain, X. Zhang, K. Huang, J. McAuley, F. Koushanfar, (2023).zPROBE: FaceSigns: Semi-Fragile Neural Watermarks for Media
Authentication and Countering Deepfakes (TOMM-2024, ACM Transactions on Multimedia Computing Communications and Applications)
-
(Trustworthy ML) Z. Ghodsi*, M, Javaheripi*, N. Sheyban*, X. Zhang*, K, Huang, & F. Koushanfar, (2023).zPROBE: Zero Peek Robustness Checks for Federated Learning, (ICCV23)
-
(Trustworthy ML) M. Samragh, S. Hussain, X. Zhang, K. Huang, & F. Koushanfar (2021). On the Application of Binary Neural Networks in Oblivious Inference. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 4630-4639)
-
(Watermarking) S. Hussain, N, Sheybani, P. Neekhara, X. Zhang, J. Duarte, F. Koushanfar (2022) FastStamp: Accelerating Neural Steganography and Digital Watermarking of Images on FPGAs. ICCAD’22)
-
(Crypto) N. Sheybani, X. Zhang, S. U. Hussain, F. Koushanfar. SenseHash: Computing on Sensor Values Mystified at the Origin. IEEE (TETC-2021)
-
(Hardware Security) H. Chen, X. Zhang, K. Huang, F. Koushanfar. “AdaTest: Reinforcement Learning and Adaptive Sampling for On-chip Hardware Trojan Detection,” ACM Transactions on Embedded Computing Systems (TECS) 2022.
-
[PATENT: FACESIGNS: SEMI-FRAGILE NEURAL WATERMARKS FOR MEDIA AUTHENTICATION AND COUNTERING DEEPFAKES] 2022 (Serial No.63/323,470.)
-
(Hardware Security) K. Huang, M.T.H. Anik, X. Zhang, and N. Karimi, “Real-Time IC Aging Prediction via On-Chip Sensors.” 2021 IEEE Computer Society Annual Symposium on VLSI (ISVLSI). IEEE, 2021
-
(Hardware Security) D. Ma, X. Zhang, et al. “DEVoT: Dynamic Delay Modeling of Functional Units under Voltage and Temperature Variations.” IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems (2021).
-
(Hardware Security) K. Huang, X. Zhang, and N. Karimi, “Real-time prediction for IC aging based on machine learning. “IEEE Transactions on Instrumentation and Measurement (TIM), vol. 68, no. 12, pp. 4756-4764, 2019.
-
[PATENT: Peasants Joy precisely pushes guiding device], CN205754440U, 2016
-
[PATENT: Bicycle lock based on bluetooth], CN205621091U , 2016
- [PATENT: Portablely lead blind waistband] CN204766395U, 2015
HIGHLIGHTED PROJECTS
[Project: Secure Retrieval-Augmented Generation on LLMs (May. 2024 - Current)]
- Implemented attack defense using a novel objective function combining adversarial loss, BERTScore, and harmful loss.
- Developed a robust RAG system for LLMs to counter universal attacks.
- Successfully prevented over 90% of state-of-the-art poisoning attacks and jailbreaking attacks on RAG-based LLMs.
[Project: Transformer-based sequence classification (Dec. 2022 - Aug. 2024)]
- Addressed the need to detect malicious activities on industry SoC log data.
- Developed transformer-based machine learning algorithms to classify hardware intrusion attacks.
- Achieved a 16% accuracy improvement in hardware intrusion detection.
Project: DNN-based media authentication and acceleration (Feb.. 2023 - Aug. 2023
)
—
- Address the growing threat of deepfakes and manipulated media, which pose significant challenges due to advances in realistic image and video synthesis techniques.
- Built a novel AI-based media authentication system using a deep learning-based semi-fragile watermarking technique.
- Achieved up to a 54% accuracy improvement with an AUC score of 0.996.
- Enabled oblivious inference in BNN.
- Explored the application of BNN in oblivious inference. Devised lightweight cryptographic protocols tailored to BNNs.
- Achieved 2x faster inference and up to 11x faster inference for binary networks.
Professional Services:
Reviewer:
- ACM Transactions on TECS
- IEEE Transactions on Dependable and Secure Computing
- The journal of supercomputing
- International Journal of Machine learning and Cybernetics
- Transactions on Machine Learning Research
- IEEE ICPR
- Transactions on Information Forensics & Security
Award
Outstanding paper award (NeurlPS 2022 TSRML)
DAC Young Fellow (58th Design Automation Conference) Nov. 2021
Honorable Mention of Mathematical Contest in Modeling Oct. 2016