Security Machine

Sadegh Farhang

917-***-**** *************@*****.***

305A Steam Services Building, University Park, PA 16802 https://sites.google.com/view/sfarhang

PROFESSIONAL SUMMARY

• More than 5 years of experience in research and development of machine learning, security, and privacy: Fairness in machine learning, machine learning model with reject option, adversarial machine learning, data-driven security, network security, usable security, censorship, security economics.

• Expert in machine learning and deep learning techniques such SVM, neural networks, convolutional neural networks, semi-supervised learning, clustering.

• Proficient in Python, TensorFlow, Keras, C/C++, R, MATLAB. EDUCATION

• Pennsylvania State University, PA

Ph.D. Informatics

GPA 3.89

Aug’15-Aug’20

• Pennsylvania State University, PA

M.Sc. Information Sciences and Technology,

GPA 3.86

Aug’15- Aug’17

• Isfahan University of Technology, Isfahan

M.Sc. Electrical Engineering,

GPA 17.99/20

• Isfahan University of Technology, Isfahan

B.Sc. Electrical Engineering,

GPA 16.62/20

Sep’11- Sep’13

Sep’07-Sep’11

PROFESSIONAL EXPERIENCE

• Autodesk Inc. May’19-Aug’19

Econometrics/Data Science Intern

Statistical models to guide the sales team to identify and prioritize new potential markets.

• USC Center for Artificial Intelligence (CAIS) May’18-Aug’18 Research Intern

Developed spatiotemporal machine learning models by using an ensemble of support vector machines (SVM) and ensemble of decision trees to predict poaching activities in wildlife. TECHNICAL SKILLS

• Advanced Mathematics: Linear algebra, matrix theory, game theory, calculus, probability, convex optimization, stochastic process, algorithms, information theory.

• Programming Skills: Python, TensorFlow, Kera, C, C++, R, MATLAB, Java, Spark, QGIS.

• Operating Systems: Linux – Ubuntu - Kali, Windows. SELECTED PROJECTS

A Bilevel Programming Approach for Machine Learning Models with Reject Option 2019-Present

• Proposing a new framework for machine learning/deep learning models with reject option for samples with high level of uncertainty using bilevel programming.

• Implementing the proposed framework in Python using TensorFlow and Keras libraries. A Framework for Analyzing Long-Term Effect of Fairness in Machine Learning 2019-Present

• Proposing a mathematical framework to model and evaluate the long-term effect of different fairness definition in machine learning.

• Implementing the proposed model in Python using TensorFlow, scikit-learn, Numpy, and Pandas. Explainable AI System for Influence Maximization Algorithms 2018-Present

• Developed a machine learning method by using decision tree, random forest, logistic regression, and support vector machine (SVM) to provide explanation to domain experts on why some members in a social network are the most influential ones.

• Implementing the proposed method in Amazon Mechanical Turk (AMT) to evaluate the usability of our proposed method.

Using Meta-Learning Approach for Learning from Imbalanced and Biased Datasets 2018-2019

• Proposed a unified data-driven regularization framework to learn from imbalanced and biased datasets and proposed an efficient algorithmic solution using stochastic gradient descent

• Implemented the proposed method and solution in Python using TensorFlow, Numpy, and Pandas libraries. Empirical Study on the Effect of Android Customization on Android-related Vulnerabilities 2018-2019

• Investigated the sources of differences in Android-related vulnerabilities in different vendors and the layers and subsystems from the Android OS affected by vulnerabilities.

• Implemented a Python web crawler using Selenium web browser to gather Android-related vulnerabilities from different websites.

Protection Assistant for Wildlife Security 2018

• Developed spatiotemporal machine learning models for prediction of wildlife poaching activities in national parks in Zimbabwe, Nigeria, and Malaysia and strategic planning for law enforcement.

• Preprocessed the data in R and implemented ensemble of decision trees and support vector machine (SVM) utilizing Python APIs including Pandas and scikit-learn.

An Economic Study of the Effect of Android Platform Fragmentation on Security Patch Updates 2017-2018

• Developed a game-theoretic approach to capture the competition among vendors in Android ecosystem and their incentive to differentiate their products and shirk on security investments.

• Developed a solution that incentivizes vendors to invest in minimum level of security while they are reducing their price.

Empirical Evidence and Game-Theoretic Approach for Time-Based Security 2016-2017

• Preprocessed and analyzed publicly available data breaches datasets to provide insights on the timing of security incidents and responses using Python APIs including Pandas and Numpy.

• Developed a game-theoretic framework that provides optimal timing for response to security breaches. Deep Learning Course Projects 2016

• Implemented a logistic regression for image classification (cat vs non-cate images) with 68% accuracy.

• Implemented a deep neural network for image classification (cat vs non-cate images) with 80% accuracy.

• Implemented a deep neural network using TensorFlow to recognize numbers from 0 to 5 in sign language.

• Implemented L2-regularization and dropout in deep neural networks to prevent overfitting. SELECTED COURSES

• Data Mining

• Artificial Intelligence

• Machine Learning and Deep Learning

• Computer and Network Security

• Information Theory

• Game Theory

• Convex Optimization

• Numerical Optimization

• Human Computer Interaction

SELECTED PUBLICATIONS

• S. Farhang, M. B. Kirdan, A. Laszka, and J. Grossklags. "An Empirical Study of Android Security Bulletins in Different Vendors." In The World Wide Web Conference (WWW) 2020.

• M. M. Kamani, S. Farhang, M. Mahdavi, and J. Wang. "Targeted Data-Driven Regularization for Out-of- Distribution Regularization." In ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) 2020.

• S. Farhang, M. B. Kirdan, A. Laszka, and J. Grossklags. "Hey Google, What Exactly Do Your Security Patches Tell Us? A Large-Scale Empirical Study on Android Patched Vulnerabilities." In Workshop on the Economics of Information Security (WEIS) 2019.

• M. M. Kamani, S. Farhang, M. Mahdavi, and J. Wang. "Targeted Meta-Learning for Critical Incident Detection in Weather Data." In Climate Change: How AI Can Help? Workshop at International Conference on Machine Learning

(ICML) 2019.

• M. Nasr, S. Farhang, A. Houmansadr, and J. Grossklags. "Enemy at the Gateways: A Game-Theoretic Approach to Proxy Distribution." In Network and Distributed System Security Symposium (NDSS) 2019.

• S. Farhang, J. Weidman, M. M. Kamani, J. Grossklags, and P. Liu. "Take It or Leave It: A Survey Study on Operating System Upgrade Practices." In Annual Computer Security Applications Conference (ACSAC) 2018.

• S. Farhang, A. Laszka, and J. Grossklags. "An Economic Study of the Effect of Android Platform Fragmentation on Security Patch Updates." In Conference on Financial Cryptography and Data Security (FC) 2018.

• S. Farhang, and J. Grossklags. "FlipLeakage: A Game-Theoretic Approach to Protect Against Stealthy Attackers in the Presence of Information Leakage." In Conference on Decision and Game Theory for Security (GameSec) 2016.

• S. Farhang, Y. Hayel, and Q. Zhu. "PHY-Layer Location Privacy-Preserving Access Point Selection Mechanism in Next-Generation Wireless Networks." In IEEE Conference on Communications and Network Security (CNS) 2015.

Contact this candidate