Parisa Kaghazgaran
Ph.D.
Department of Computer Science and Engineering
Texas A&M University
Office:408D, Bright Building (HRBB)
Email: kaghazgaran@tamu.edu
Google Scholar
My name is Parisa Kaghazgaran (I know how hard is to pronounce my last name! you can speak it as kagazgaran- g as in girl), a PhD student and Research Assistant in the the Department of Computer Science and Engineering at Texas A&M University. I'm working in Infolab under supervision of Dr. James Caverlee. I recieved my bachlor and master's degree from Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran.
My research interest lies in social media data mining, graph mining and applied machine learning/deep learning, with special focus on misbehavior and fake reviews detection. Currently, I am doing research on generative models and transfer learning for fake review generation. I have done extensive research in information security with the focus on privacy preserving protocols before joining Infolab.

  • Parisa. Kaghazgaran, J. Caverlee, A. Squicciarini "Combating Crowdsourced Review Manipulators: A Neighborhood-Based Approach". ACM WSDM, 2018, (Acceptance Rate 16%).[pdf] [slides] [Poster]
  • Majid Alfifi, Parisa. Kaghazgaran, James Caverlee, Fred Morstatter "Measuring the Impact of ISIS Social Media Strategy", MIS2 Workshop (in conjunction with WSDM), 2018.[pdf] [poster]
  • Parisa. Kaghazgaran, J. Caverlee, M. Alfifi "Behavioral Analysis of Review Fraud: Linking Malicious Crowdsourcing to Amazon and Beyond", ICWSM, 2017, (Acceptance Rate 19%) (short paper).[pdf] [poster]
  • Majid Alfifi, James Caverlee, Parisa Kaghazgaran, Fred Morstatter "Collective Influence and Behavioral Analysis of ISIS Social Media Strategy", ICWSM 2019, (Acceptance Rate 16%).
  • Shanshan Li, James Caverlee, Wei Niu , Parisa Kaghazgaran "Crowdsourced App Review Manipulation", SIGIR, 2017, (Acceptance Rate 30%) (short paper). [pdf] [poster]
  • Parisa Kaghazgaran, Hassan Takabi, Flannery. H. C, Armando. S "Communication-efficient Private Identification Based on Oblivious Transfer Extensions", will appear in Elsevier Computer & Security Journal, 2018 (IF: 2.65).
  • Salman Niksefat*, Parisa Kaghazgaran*, B. Sadeghiyan, "Privacy Issues in Intrusion Detection Systems: a Taxonomy, Survey and Future Directions", Elsevior Computer Science Review Journal, 2017. (SNIP:5.8, SJR: 1.678) [pdf]
  • Parisa Kaghazgaran, H. Takabi, "Privacy Preserving EEG-based Authentication", WiCyS, 2016
  • Parisa Kaghazgaran, H. Takabi, "Differentially Private Decision Tree Learning from Distributed Data", IEEE Symposium on Privacy & Security (S&P), 2015
  • Parisa Kaghazgaran, H. Takabi, "Towards a New Insider Detection Framework", Grace Hopper, 2015
  • Reviewer at ICWSM 2019
  • Program Committee of MIS2 workshop
  • Journal Reviewer: International Journal of Cooperative Information Systems
  • External Reviewer: WSDM'(17, 18), WWW'(17, 18), SIGIR'17, RecSys'17, ASONAM'17, CIKM'17
  • Invited Talk: Misinformation and Misbehavior Mining on the Web (MIS2'18 Workshop)
  • Research Intern, Microsoft AI & Research (MSR), Jun. 2018-Aug. 2018
    Project: Neural Ranking Model for Ad-hoc Retrieval from Search Click-through Data
    Area: Deep learning, Ranking, Search click-through modeling
  • Neural Networks and Deep Learning
  • Structuring Machine Learning Projects
  • Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization
  • Nominated by the CSE Department (Texas A&M University) for 2019 Microsoft Research PhD Fellowship.
  • ACM SIGIR Student Travel Grant, 2018
  • Recognized as Outstanding CSE Ph.D. Student at UNT, 2016
  • Tapia Conference Scholarship, 2016
  • Women in Cyber Security (WiCyS) Conference Scholarship, 2016
  • Computing Research Association Women (CRA-W) Scholarship, 2016
  • IEEE Symposium on Security and Privacy Travel Grant, 2015
  • Department Travel Award for Grace Hopper Celebration, 2015, 2017, 2018
  • Graduate Assistantship Tuition Scholarship Award (GATS) from UNT for 2014/2015 AY
  • Ranked as the Top master student according to GPA, Amirkabir University of Technology (Tehran Polytechnic)
  • Exceptional Talent Award for master Program Amirkabir University of Technology (Tehran Polytechnic)
  • Achieving the 301th rank in the National University Entrance Examination for the bachlor degree among more than 400,000 participants
  • Reliability evaluation of reviews on Amazon, Yelp and App store
  • This project evaluates rating patterns to identify manipulation attacks using traditional machine learning features and RRN model. Check out our (Demo)
  • Surpassing the limit of collecting data from Twitter
  • This project implements four different algorithms to improve Twitter samples coverage by clustering co-occurring words.
  • Crowd-sourcing websites and Amazon Reviews Crawlers
  • This project focuses on collecting actual evidence of review manipulation by developing web crawlers. Crawlers are implemented in python using Beautiful Soup library. The output of this project is a dataset of fake reviews (it will be made public soon)
  • Behavioral analysis of fraudulent and regular users on Amazon
  • This project investigates reviewers behaviors like rating, review burstiness and language features using statistical and NLP techniques e.g., cross entropy
  • Network Behavior of Reviewers in Amazon
  • This project propagates suspiciousness by exploiting behavioral and social features to identify "similar" users through a random walk over a "suspiciousness" graph and uncovers (hidden) distant users who serve structurally similar roles by mapping users into a low-dimensional embedding space that captures community structure.
  • Text-based mini search engine
  • This project follows three main learning objectives: 1. The basics of tokenization and its effect on information retrieval 2. The basics of index building and Boolean retrieval 3. The basics of the Vector Space model and ranked retrieval
  • Link Analysis, PageRank and TrustRank Algorithms
  • This project explores real-world challenges of building a graph (e.g., twitter following graph) and implements/evaluates PageRank over this graph. As part of this project we also investigate factors that impact a page's rank on Google and Bing.
  • Naive-Bayes and Rocchio Classifiers on Yelp review data
  • This project aims to build and evaluate ML classifiers from scratch in python.
  • Sentiment Classification of tweets based on their sad or happy emojis
  • Privacy Preserving Distance Calculation Among Distributed Samples. [Code]
  • This project follows security two-party computation scenarios in which two distant parties wish to compute a function on their joint data but they do not will to reveal the data due to privacy constraints. A general framework for privacy preserving distance calculation is developed using socket programming to communicated between server and client while Oblivious Transfer protocol protects data privacy. This will appear in Elsevier Computer and Security Journal.
Last Updated: Sep. 2018