About Me
I am a Postdoctoral Researcher working on Privacy-Preserving Machine Learning at Information Processing and Communications Lab at Imperial College London with Prof. Deniz Gunduz. I have completed my PhD in Computer Science at Queen Mary University of London as a member of Centre for Intelligent Sensing under the supervision of Prof. Andrea Cavallaro (formerly, Dr. Hamed Haddadi) and Dr. Richard G. Clegg, supported by Life Sciences Institute. My Ph.D. research was focused on privacy-preserving machine learning mechanisms for behavioural data analysis, particularly for motion sensor data generated by the mobile and wearable devices. I have been a Ph.D. intern at Brave Software where I worked on privacy-preserving content personalization, working with Dr. Dimitrios Athanasakis and Dr. Benjamin Livshits. I have also been a visiting PhD student at Imperial College London, as a member of DataboxProject, working with Dr. Hamed Haddadi. Please download my CV for more information.
Preprints
◉ Quantifying Information Leakage from Gradients
[arXiv] [ICLR Workshop Version]Fan Mo, Anastasia Borovykh, Mohammad Malekzadeh, Hamed Haddadi, Soteris
Demetriou
• 2021 - Preprint (Under Review).
• The preliminary version has been presented at ICLR Workshop on Distributed and Private Machine Learning
(DPML-2021).
◉ Running Neural Network Inference on the NIC
[arXiv].Giuseppe Siracusano, Salvator Galea, Davide Sanvit, Mohammad Malekzadeh, Hamed
Haddadi, Gianni Antichi, Roberto Bifulco.
• 2020 - Preprint (Under Review).
Publications
◉ Honest-but-Curious Nets: Sensitive Attributes of Private Inputs can be Secretly Coded into the Entropy of Classifiers' Outputs
[arXiv] [Code] Mohammad Malekzadeh, Anastasia Borovykh, Deniz Gündüz.
• 2021 - Accepted for the ACM Conference on Computer and Communications Security (CCS) (Under Minor Revison).
◉ DANA: Dimension-Adaptive Neural Architecture for Multivariate Sensor Data
[arXiv] [Code] [Tutorial]Mohammad Malekzadeh, Richard G. Clegg, Andrea Cavallaro, and Hamed Haddadi.
• 2021 - ACM Journal on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT)
• Will also be presented at ACM conference for Ubiquitous Computing (UbiComp 2021)
◉ Dopamine: Differentially Private Federated Learning on Medical Data
[PDF] [Code]Mohammad Malekzadeh, Burak Hasircioglu, Nitish Mital, Kunal Katarya, Mehmet Emre Ozfatura, Deniz Gündüz
• 2021 - The Second AAAI Workshop on Privacy-Preserving Artificial Intelligence
(PPAI-21)
• Selected as the best submission to the ITU AI/ML in 5G Challenge, sub-challenge Privacy Preserving AI/ML in 5G networks for healthcare applications, ITU-ML5G-PS-022
◉ Privacy and Utility Preserving Sensor-Data Transformations
[Journal PDF]. [arXiv]. [Code]Mohammad Malekzadeh, Richard G. Clegg, Andrea Cavallaro, and Hamed Haddadi.
• 2020 - Pervasive and Mobile Computing Journal, Elsevier.
◉ Privacy-Preserving Bandits
[Proceedings PDF]. [arXiv]. [Code]Mohammad Malekzadeh, Dimitrios Athanasakis, Hamed Haddadi, and Benjamin Livshits
• 2020 - Conference on Machine Learning and Systems (MLSys), USA.
◉ Modeling and Forecasting Art Movements with CGANs
[Journal PDF]. [arXiv]. [Code]Edoardo Lisi, Mohammad Malekzadeh, Hamed Haddadi, F. Din-Houn Lau, Seth Flaxman.
• 2020 - Royal Society Open Science Journal, The Royal Society.
◉ Mobile Sensor Data Anonymization
[Proceedings PDF]. [arXiv]. [Code]Mohammad Malekzadeh, Richard G. Clegg, Andrea Cavallaro, and Hamed Haddadi.
• 2019 - ACM/IEEE International Conference on Internet-of-Things Design
and Implementation (IoTDI), Canada.
◉ Protecting Sensory Data against Sensitive Inferences
[Proceedings PDF]. [arXiv]. [Code]Mohammad Malekzadeh, Richard G. Clegg, Andrea Cavallaro, and Hamed Haddadi.
• 2018 - In Proceedings of the 1st EuroSys Workshop on Privacy by Design
in
Distributed Systems (W-P2DS), Portugal.
◉ Replacement AutoEncoder: A Privacy-Preserving Algorithm for Sensory Data Analysis
[Proceedings PDF]. [arXiv]. [Code]Mohammad Malekzadeh, Richard G. Clegg, and Hamed Haddadi.
• 2018 - ACM/IEEE International Conference on
Internet-of-Things Design and Implementation (IoTDI), USA.
◉ Technical Report : Fairness in Algorithmic Decision Making
[PDF] Paul-Marie Carfantan, Omar Costilla-Reyes, Delia Fuhrmann, Jonas Glesaaen, Qi
He, Andreas Kirsch, Julie Lee, Mohammad Malekzadeh, Esben Sørig, Caryn Tan, Emily Turner;
Vinh Quang.
• 2018 - Data Study Group team. (2019, February 5). Data Study Group Final
Report: Accenture. Zenodo.
◉ Gamified Incentives: A Badge Recommendation Model to Improve User Engagement in Social Networking Websites
[Journal PDF].[Code] Reza Gharibi and Mohammad Malekzadeh.
• 2017 - International Journal of Advanced Computer Science and
Applications (IJACSA).
◉ A Multi-Generational Social Learning Model: the Effect of Information Cascade on Aggregate Welfare
[Proceedings PDF]Marziyeh Barghandan, Mohammad Malekzadeh, Atefeh Safdel and Iren Mazloomzadeh.
• 2014 - The IEEE/ACM International Conference on Advances in Social
Networks Analysis and Mining, ASONAM 2014, China.
◉ Social Balance and Signed Network Formation Games
[PDF]Mohammad Malekzadeh, MohammadAmin Fazli, Pooya Jalaly Khalilabadi, Hamid R.
Rabiee, MohammadAli Safari.
• 2011 - The 5th SNA-KDD Workshop ’11 (SNA-KDD’11), USA.
◉ Persian CAPTCHA system to prevent automatic subscribing of software robots in web pages
[PDF](Persian)Mohammad Malekzadeh and Mehdy Bohlool.
• 2008 - The 13th National CSI Computer Science , Iran.