Foto von Marvin Xhemrishi

M.Sc. Marvin Xhemrishi

Technische Universität München

Professur für Codierung und Kryptographie (Prof. Wachter-Zeh)

Postadresse

Postal:
Theresienstr. 90
80333 München

Abschlussarbeiten

Angebotene Abschlussarbeiten

Private and Efficient Vertical Federated Learning

Beschreibung

Federated Learning (FL) tackles the problem of learning from decentralized data from different entities called clients. In vertical FL, the clients have the same data samples (entries) but different features. However, during the learning, private data can be leaked. In [1] the authors suggest a private VFL framework that uses homomorphic encryption and multi-party computation. Another problem that can arise is that the clients might dropout during the learning. In [2] the authors tackle the privacy and the straggler resiliency problem. This seminar aims to understand VFL settings using privacy-preserving and straggler-resiliency techniques.

References:

[1] - https://arxiv.org/abs/2008.06170

[2] - https://arxiv.org/abs/2304.13407

Kontakt

marvin.xhemrishi@tum.de

Betreuer:

Differentially-Private and Robust Federated Learning

Beschreibung

Federated learning is a machine learning paradigm that aims to learn collaboratively from decentralized private data owned by entities referred to as clients. However, due to its decentralized nature, federated learning is susceptible to poisoning attacks, where malicious clients try to corrupt the learning process by modifying their data or local model updates. Moreover, the updates sent by the clients might leak information about the private data involved in the learning. This thesis aims to investigate and combine existing robust aggregation techniques in FL with differential privacy techniques.

References:

[1] - https://arxiv.org/pdf/2304.09762.pdf

[2] - https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9757841

[3] - https://dl.acm.org/doi/abs/10.1145/3465084.3467919

Voraussetzungen

- Knowledge about machine learning and gradient descent optimization

- Proficiency in Python and PyTorch

- Undergraduate statistics courses

- Prior knowledge about differential privacy is a plus

Kontakt

marvin.xhemrishi@tum.de

luis.massny@tum.de

Betreuer:

Laufende Abschlussarbeiten

Secure Federated Learning with Differential Privacy

Beschreibung

Federated learning is a machine learning paradigm that aims to learn collaboratively from decentralized private data owned by entities referred to as clients. However, due to its decentralized nature, federated learning is susceptible to model poisoning attacks, where malicious clients try to corrupt the learning process by modifying local model updates. Moreover, the updates sent by the clients might leak information about the private data involved in the learning. The goal of this work is to investigate and combine existing robust aggregation techniques in FL with differential privacy techniques.

References:

[1] - https://arxiv.org/pdf/2304.09762.pdf

[2] - https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9757841

[3] - https://dl.acm.org/doi/abs/10.1145/3465084.3467919

Voraussetzungen

- Basic knowledge about machine learning and gradient descent optimization

- First experience with machine learning in python

- Undergraduate statistics courses

- Prior knowledge about differential privacy is a plus

Betreuer:

Contribution scoring in Federated Learning

Beschreibung

Federated learning (FL) is a machine learning paradigm that aims to learn collaboratively from decentralized private data, owned by entities referred to as clients. In real-world applications of FL it is important to score the contribution of each client. The goal of this seminar is to provide a high-level overview of existing contribution-scoring techniques in federated learning using [1-2] and other references. 

Voraussetzungen

References: 

[1] - https://ieeexplore.ieee.org/document/10138056

[2] - https://arxiv.org/pdf/2403.07151.pdf

Betreuer:

Implementation of model poisoning attacks in federated learning

Beschreibung

Federated learning is a machine learning paradigm where decentralized entities (clients) collaboratively learn using their private data. A central server acts as a coordinator of the learning process. Due to the sensitivity of the private data involved,  the data cannot be transferred. A salient problem of federated learning is the presence of malicious clients, which are clients that try to destroy the learning process. Malicious clients can do this by corrupting their data and/or by modifying their local model updates. The goal of this project is to understand how model poisoning attacks and defense strategies perform under different scenarios of federated learning using experiments. 

References: 

[1]- https://www.ndss-symposium.org/wp-content/uploads/ndss2021_6C-3_24498_paper.pdf

[2]- https://arxiv.org/pdf/1903.03936.pdf

[3]- https://arxiv.org/pdf/2304.00160.pdf

 

Voraussetzungen

  • Basic knowledge of machine learning
  • Python programming skills, knowledge of PyTorch is an advantage 

Kontakt

marvin.xhemrishi@tum.de

Betreuer:

Publikationen

2022

  • Bitar, R.; Xhemrishi M.; Wachter-Zeh A.: Adaptive private distributed matrix multiplication. IEEE Transactions on Information Theory, 2022 mehr…
  • Garb K.; Xhemrishi M.; Kürzinger L.; Frisch C.: The Wiretap Channel for Capacitive PUF-Based Security Enclosures. IACR Transactions on Cryptographic Hardware and Embedded Systems, 2022 mehr…
  • Maringer G.; Xhemrishi M.; Puchinger S.; Garb K.; Liu H.; Jerkovits T.; Kürzinger L.; Hiller M.; Wachter-Zeh A.: Analysis of Communication Channels Related to Physical Unclonable Functions. Workshop on Coding and Cryptography (WCC), 2022 mehr…
  • Xhemrishi M.: Distributed Linear Computations over Private Sparse Matrices. IEEE European School of Information Theory, 2022 mehr…
  • Xhemrishi M.: Distributed Linear Computations over Private Sparse Matrices. Munich Workshop on Coding and Cryptography, 2022 mehr…
  • Xhemrishi M.: Computational Code-Based Privacy for Coded Federated Learning. TUM ICE Workshop Raitenhaslach, 2022 mehr…
  • Xhemrishi M.; Bitar R.; Wachter-Zeh A.: Distributed Matrix-Vector Multiplication with Sparsity and Privacy Guarantees. IEEE International Symposium on Information Theory, 2022 mehr…
  • Xhemrishi M.; Egger M. Bitar R.: Efficient Private Storage of Sparse Machine Learning Data. IEEE Information Theory Workshop, 2022 mehr…
  • Xhemrishi M.; Graell i Amat A.; Rosnes E.; Wachter-Zeh A.: Computational Code-Based Privacy in Coded Federated Learning. IEEE International Symposium on Information Theory, 2022 mehr…

2021

  • Maringer G.; Xhemrishi M.; Puchinger S.; Garb K.; Liu H.; Jerkovits T.; Hiller M.; Wachter-Zeh A.Kürzinger L.;: Analysis of Communication Channels related to Physically Unclonable Functions. arXiv, 2021 mehr…
  • Xhemrishi M.: Trade-off between privacy and sparsity in coded computing. 36th meeting of ITG Professional Group "Applied Information Theory", 2021 mehr…

2020

  • Bitar R.; Xhemrishi M.; Wachter-Zeh A.: Rateless Private Matrix-Matrix Multiplication. European School of Information Theory, 2020 mehr…
  • Bitar R.; Xhemrishi M.; Wachter-Zeh A.: Fountain codes for Private Distributed Matrix-Matrix Multiplication. International Symposium in Information Theory and its Applications, 2020 mehr…

2019

  • Xhemrishi, M.; Coşkun, M. C.; Liva, G.; Östman, J.; Durisi, G. : List Decoding of Short Codes for Communication over Unknown Fading Channels. , Workshop on Coding, Cooperation, and Security in Modern Communication Networks (COCO 2019) . , 2019 mehr…
  • Xhemrishi, M.; Coşkun, M. C.; Liva, G.: List Decoding for Fading Channels. Oberpfaffenhofen Workshop on High Throughput Coding (OWHTC) , 2019 mehr…
  • Xhemrishi, M.; Coşkun, M. C.; Liva, G.; Östman, J.; Durisi, G.: List Decoding of Short Codes for Communication over Unknown Fading Channels. Asilomar Conference on Signals, Systems, and Computers, 2019 mehr…