Metadata
Title
Theses
Category
general
UUID
fd24e061649f4a128982d6724d9d9946
Source URL
https://disco.ethz.ch/theses
Parent URL
https://inf.ethz.ch/studies/semester-and-master-theses.html
Crawl Time
2026-03-09T07:13:32+00:00
Rendered Raw Markdown

Theses

Source: https://disco.ethz.ch/theses Parent: https://inf.ethz.ch/studies/semester-and-master-theses.html

Our group has several open research projects that are suitable for a master (M), bachelor (B) or semester (S) theses. In addition, we have group projects, and labs for distributed systems majors; projects marked as B or S are usually also suitable for labs or group projects.

If you have an idea for a project that is not on our list but falls into our , we would be excited to hear about it.
Available Theses & Labs
Title Type Contact/Supervisor Assigned
Entanglement Optimization for Quantum Mobile Networks B/M/S available
Provably Robust Cooperative Multi-Agent RL M available
An Investigation into the Limits of Tiny Recursive Model Reasoning M/S , Giacomo Camposampiero available
Attacking P2P networks M/S available
Bring Your Own Algorithm M/S Andrei Constantinescu available
Can LLMs Forecast Swiss Referendum Outcomes? M/S available
Consensus: Fast and Lite M/S , available
Cool-Chic MoE: Fast and Expressive Image Decoding M/S available
Distributed Reinforcement Learning for Optimization (ETH-MIT Joint Project) M/S available
Federated Dense Associative Networks M available
Federated Neural ODEs for Efficient Distributed Physical Modeling B/M/S available
Karma, Fair and Square? M/S available
Label-Free Compositional Generalization M/S Giacomo Camposampiero, available
Measuring Task Similarity for Unbiased Benchmark Aggregation M available
Optimal Transport for Federated Learning: Bridging Semantics and Alignment through Mechanistic Interpretability M , available
Resilient Backbone Deployment under Cooperative Adversaries M/S available
Resource-Aware Adaptive Split Learning with Reinforcement Learning B/M/S available
Sparse Autoencoders for Sensitive Feature Removal in Federated Learning M , available
Topics in Computational Voting and Game Theory M/S Andrei Constantinescu available
Topics in Distributed and Decentralized Artificial Intelligence B/M/S available
Topics in Game Theory M/S available
Topics on Distributed Reinforcement Learning M/S available
Current Theses & Labs
Title Type Contact/Supervisor Assigned Student(s)
Adaptive and Resilient Resource-Aware Split Learning S FS 26 Ahmed Ibrahim
Adversarial and Adaptive Data-driven Cooperative PHY Attacks and Countermeasures for Mobile Communications B FS 26 Rahul Rahman
Alignment Optimization between SAE through Monosemanticity B , FS 26 Mohamed Kamel Abdelrahman
Byzantine Micro-Split Learning S FS 26 Efe Berke Alici
Communication Complexity in Unreliable Networks S , FS 26 Lucca Guerin Saturnino
Data-driven Coordination for Decentralized Federated Floating Radio Maps in MANETs B FS 26 Heiko Graef
Differentiable Logic Gate Networks with Classic Computer Vision Encoding S , FS 26 Leonardo Puzzuoli
Diffusion model guidance S FS 26 David Oancea
Enhancing Optimization for Label-Free Federated Compositional Learning S , Giacomo Camposampiero FS 26 Sarah Kuhn
Enhancing Privacy of Communication-Efficient Policy-Exchange Processes in Multi-Environment Federated Reinforcement Learning S FS 26 Radu Cretu
Event-based data classification with Differentiable Logic Gate Networks S , FS 26 Luis Clavuot Bapst
Federated Difflogic B FS 26 Felix Kunze
Federated Hopfield Networks B FS 26 Afsaneh Bababak
Federated Learning with Adaptive Quantization Using Optimal Transport and Weight Importance S , FS 26 Till Dubach
Federated Multi-Agent Reinforcement Learning with Low-Latency Logical Policies under Heterogeneous Environments B FS 26 Lars Tim Hulsbergen
Federated Neural Architecture Search B FS 26 Rafael Bragança Oliveira
Federated Neural Controlled Differential Equations for Distributed Physical Modeling B FS 26 Noah Kuhn
Federated Safety Heads for Prompt Injection Detection for Agentic LLMs S FS 26 Octavian Sandu
Federated Semi-Supervised Learning with Prototype-based Label Propagation using Iterative Graph Refinement B FS 26 Laszlo Doriat
Incentive Protocols for Decentralized Federated Learning in High-Churn Networks B FS 26 Jan Cerny
Jet Substructure Classification with TorchLogic B , FS 26 Adam Amanbaev
Label-Free Compositional Generalization B Giacomo Camposampiero, FS 26 Simon Arnold
LLMs and Political Semantics S FS 26 Nassim Beladel
Measuring Decentralization in DAOs B FS 26 Sven Bruhin
Normalized Vision Transformers S , FS 26 Ivanovas Anselm
On the Role of Encoder and Decoder Expressiveness in SAE Feature Alignment S , FS 26 Nicolas Schluep
Optimal Transport for Model Merging under the Linear Representation Hypothesis B , FS 26 Leon Zitzelsberger
Physical- and Media-Access-Control-Layer Optimization in Quantum Mobile Networks S FS 26 Tim Joller
Probabilistic Attribute Invariant Networks and Compositionally-Weighted Conformal Prediciton for Uncertainty Quantification in Compositional Learning S Giacomo Camposampiero, FS 26 Tamer van Teeffelen
Recursive Reasoning in Multi-Agent Reinforcement Learning B FS 26 Alex David Lanz
Resilient Communication-Optimal State Consensus in Jammed MANETs B FS 26 Arwin Baur
Robust Multi-Agent Adversarial Policies in Jammed MANETs S FS 26 Mikolaj Stecko
Robust Yield Curve Approximation under Sparse Retraining and Non-Stationary Market Conditions S FS 26 Marko Babovic
Solving Minesweeper with Differentiable Logic Gate Networks S , FS 26 Arash Bayat
Splittable Neural Architecture Search B FS 26 Tim Oppliger
Towards Gödel Machines: Practical Self-Referential LLM Systems S , FS 26 Severin Bratus
Advancing Neural Reasoning S HS 25 Michal Tesnar
Automated Foosball Commentary B , HS 25 Victor Willems
Byte Latent Transformers for Audio S , HS 25 Nandor Kofarago
Cinematic Source Separation S , HS 25 Atoof Shakir
Codec Translation S , HS 25 Leo Schmidt-Traub
Cool-chic M HS 25 Jakub Parada
Detecting Blood Vessel Networks from 3D Scans S , HS 25 Sageder Clemens
DiffLUT M , HS 25 Simon Bührer
Embedding Political Opinions S HS 25 Johanna Dciofalo
Federated Multi-Agent Reinforcement Learning B HS 25 Peter Parsons
M HS 25 Enrico Mann
Learning Structural Representations in Graphs S HS 25 Shawn Shawn Lim
Linear Depth Graph Models S , HS 25 Basil Rohner
Management of a Decentralized Storage Network M , HS 25 Fabian Schaich
Market Efficiency in Information Market B HS 25 Berdan Deniz
Meta-Federated Reinforcement Learning: Learning to Learn from Distributed Expert Knowledge S HS 25 Adam Suma
Multi-Agent Collaboration for Safe and Robust RLHF B HS 25 Valentin Graf
Optimization of Multi-Agent Team Composition and Communication for Enhancing Reliability in Industrial Applications M HS 25 Atien Kosin
Practical Federated RL: Implementation and Interactive Scenario Design with Flower.ai B HS 25 Joel Maag
Pretraining GNNs S , HS 25 Sebastian Brunner
Questions Similarity in Voting Advice Applications B , HS 25 Liv Weiss
Recursive Reasoning S HS 25 Peer Rheinboldt
Scalable Awareness for Resilient Cooperative Multi-Agent MANET Deployment against Jamming B HS 25 Alex Ferdinand Niederfriniger
Tackling Multi-Agent Policy-Gradient Variance with Stochastic Variance-Reduced Estimators M HS 25 Vili Kosonen
Topics on distributed stable matching S HS 25 Michael Heider
Trust and Resilience in Distributed Reinforcement Learning B HS 25 Pascal Gisiger
Unleashing Forward Thinking in Language Models M , HS 25 Aleksander Lorenc
Bounds on the Coding Gap for Broadcast S , Andrei Constantinescu FS 25 Natalie Suter
Distributed Graph Problems with Malicious Actors M , FS 25 Lucas Weitzendorf
Interpreting the Reversal Curse of LLMs M FS 25 Giovan Battista Landolina
Reasoning Benchmark for LLMs B , FS 25 Fabian Farestam
Sample Efficiency in Text Diffusion Pretraining S FS 25 Ahmad Fraij
Scalar Quantization for Audio Compression S , FS 25 Fei Gao
Scaling laws for test-time compute with SID.ai M FS 25 Jonathan Manz
View-Specific Video Compression M FS 25 Niklas Pohl
Game Theory for Games: Algorithms and Axioms G Andrei Constantinescu FS 24 Mihail Lavric
Learning from Demonstrations G FS 24 Jonas Pai
Past Theses & Labs
Title Type Contact/Supervisor Assigned Student(s)
Homo Moralis: Mixed Strategy Equilibria in Sequential Games [confidential] B HS 25 Pirmin Ballmer
A Framework for Grounding LLM Social Simulations: A Pilot Study on LLM Depolarisation and Calibration [confidential] S HS 25 Mabel Wylie
Forecasting Swiss Federal Referendums: A Two-Tower Architecture Using LLM-Extracted Features from Official Voting Texts [confidential] S HS 25 Stefan Gerber
Exploring Normalized Transformers [confidential] M , HS 25 Shakir Yousefi
Splitting Queries Across Engines: A Critical-Path Approach [confidential] M HS 25 Soufiane Barrada
Integrating AI Agents with Blockchain Gaming Economies [confidential] G HS 25 Samuel Huber
Can AI Agents Agree? [confidential] B HS 25 Leonardo Rugli
Detecting Blood Vessel Networks from 3D Scans [confidential] S , HS 25 Florian Portmann
Variable Hardness Training Instance Generation for Neural SAT Solvers [confidential] S , HS 25 Frederik van der Brugge
Subliminal Signals in Preference Labels [confidential] G , HS 25 Isotta Magistrali
Eigenvector-Masked Autoencoders B , FS 25 Anja Buchmann
Aggregation Functions for Two-Team Games [confidential] M Andrei Constantinescu, FS 25 Georgia Hadjipanayiotou
AlphaZero meets ARC: Explorative Capability Acquisition through Goal-Conditioned Adversarial Curriculum Creation [confidential] M , FS 25 David Jenny
A Survey on the Abstraction and Reasoning Corpus (2019–2025) [confidential] S FS 25 Severin Bratus
α-Asynchronous Updates in Graph Neural Networks [confidential] B FS 25 Tim Krappel
Beyond Deepfake Detection [confidential] B , FS 25 Antonis Asonitis-Spyridon
Context-aware In-Car Recommender System [confidential] M FS 25 Andri Furrer
Alignment-Aware Decoding [confidential] M , FS 25 René Cáky
Empirical Evidence of the Mangrove State Replication Protocol for Blockchains [confidential] B , FS 25 Cedric Schwyter
Data Attribution via Bidirectional Gradient Optimization [confidential] B , FS 25 Kaan Bayraktar
Deep Differentiable Logic Gate Networks: Neuron Collapse Through a Neural Architecture Search Perspective [confidential] S , FS 25 Shakir Yousefi
A Comprehensive Benchmark for Graph-Based Abstract Reasoning [confidential] M , FS 25 August Rønberg
Recurrent Deep Differentiable Logic Gate Networks [confidential] S , FS 25 Simon Bührer
Effectiveness of Multi-Scale Aggregation for Adversarial Robustness B , FS 25 Emerson Leonardo Azevedo Aguiar
Empirical Analysis of Wash Trading in Celo’s Stablecoin Ecosystem [confidential] B , FS 25 Daniel Hunziker
GNN-enhanced AlphaZero for Two-player Games S , FS 25 Pacurar Andrei-Horia
Evaluating AI-Generated Image Detection Across Resolution and Complexity [confidential] B , FS 25 Lorenzo Alessandro Vettor
PoliSpace: Representing Political Questions and Opinions in a Shared Embedding Space [confidential] M FS 25 Johanna Dciofalo
Adversarial Attacks on Machine Learning S FS 25 Nastasia Morel
Creating Fair Foosball Match Schedules [confidential] G , Andrei Constantinescu FS 25 Julia Hütte
Bias Research on Generated Music from around the World [confidential] M , FS 25 Ahmet Solak
Faster DiffLogic [confidential] S , FS 25 Joshua Durrant
Geometry-Aware Graph Generation with Riemannian Diffusion and Local Expansion [confidential] M , FS 25 Anisha Mohamed Sahabdeen
The Impact of Training Data on Adversarial Examples [confidential] B , FS 25 Marco Zimmerli
Light Differentiable Logic Gate Networks [confidential] S , FS 25 Lukas Rüttgers
Scalability and Expressiveness of the Group-Sum Layer in Differentiable Logic Gate Networks [confidential] S , FS 25 Sven Brändle
Failing at Basic Visual Perception but Solving Complex Visual Tests? Evaluating Fundamental Visual Perception Skills of VLMs on Synthetic Datasets [confidential] B , FS 25 Elias Hug
Benchmarking LLM Sampling Strategies [confidential] B , FS 25 Hugo Posada Saiz
MaskBeat: Drum–Chord Sequence Generation via Theory-Aware Losses [confidential] B , FS 25 Karim Galal
Pandora’s Box of Solana: MEV in the Shadows of Private Mempools [confidential] M FS 25 Albin Mamuti
To Bid or Not to Bid? The Goods, the Chores and the Karma [confidential] S FS 25 Cécile Michel
From E-Commerce to Editorials: Garment Retrieval for Virtual Photo-Shoot Applications [confidential] M , FS 25 Yannick Hauri
EuroSpeech: A Large-Scale Multilingual Speech Corpus from Parliamentary Recordings via Scalable Open-Source Pipelines [confidential] B , FS 25 Samuel Pfisterer
Multimodal Graph-Language Models [confidential] S , FS 25 Simon Schlude
Fine-tuning Data Extraction from Large Language Models [confidential] M , FS 25 Samuel Räber
Parametric Neural Amp Modeling with Active Learning [confidential] S , FS 25 Longxiang Jiao
Ranking Movies with LLMs B , FS 25 Tilman Sieber
A Dual Study on Analyzing Image Recompression Effects and Realism Assessment via Kolmogorov Complexity [confidential] B FS 25 Elif Özsoy
Source Separation for A Cappella Music [confidential] S , FS 25 Constantin Pinkl
Estimating Room Impulse Response using GNNs [confidential] M , FS 25 Hyun-Min Chang
rStar-ARC: Process-Preference–Guided MCTS for Stepwise Program Synthesis [confidential] B , FS 25 Pierre Roth
Scaleable Proof-of-Personhood: Exploring the Potential of PoP Combinations B FS 25 Jan Boog
Simulating Democracy: Technical Feasibility and Policy Perspectives [confidential] M , FS 25 Yann Billeter
Benchmarking Diarization Models [confidential] B , FS 25 Cesare Blaser
Steering Pretrained Drafters during Speculative Decoding [confidential] B , FS 25 Peer Rheinboldt
Synthetic Data Augmentation in Medical Image Classification: How Architecture Determines Utility B , FS 25 Lars Ruschak
Text-to-Scene using Large Reasoning Models [confidential] S , FS 25 Nick Tuninga
Mixture of Recursion and Experts [confidential] M , FS 25 Shiduo Xin
SALSA-V: Shortcut-Augmented Long-form Synchronized Audio from Videos [confidential] M , FS 25 Amir Dellali
CAST: Conditional Adaption with Soft Tokens [confidential] M HS 24 Leo Neubecker
Rethinking Adversarial Examples [confidential] M HS 24 Yahya Jabary
Yield Curve Approximation and Prediction with Recurrent Neural Networks [confidential] B Daniela Brauckhoff, HS 24 Marko Babovic
Anomaly Detection for Automated Visual Inspection [confidential] M HS 24 Mateo Diaz-Bone
Computing with Asynchronous Cellular Automata [confidential] M , HS 24 Ivan Baburin
Free-form Audio Editing using Natural Language Instructions [confidential] M , HS 24 Michael Ungersböck
Imperceptible Audio Watermarking [confidential] M , HS 24 Kyle Fearne
Automated Visual Foosball Tracking B , HS 24 Linus Baumberger
Breaking Image CAPTCHAs and designing a new CAPTCHA [confidential] M HS 24 Turlan Kuzhagaliyev
Consensus Bugs and Where To Find Them [confidential] M HS 24 Ludovic Burnier
DAO Governance Aggregation: A Comprehensive Analysis of Decentralized Governance Systems [confidential] B HS 24 Mattia Pepin Taiana
Benchmarking DeepEye [confidential] B HS 24 Ruben Widmer
EEG Benchmarking Platform [confidential] B HS 24 Yun-Chiang Feng
Efficient Real-Time Egocentric Action Recognition M HS 24 Marco Calzavara
Leveraging Pre-Trained LMs for Rapid and Accurate Structure Elucidation from 2D NMR Data [confidential] S , HS 24 Susanna Di Vita
An Empirical Analysis of Polymarket [confidential] B , HS 24 Dominique Buob
FastFeedForward Networks in Vision Transformers [confidential] S HS 24 Amaru Caceres Arroyo
Expressive GNNs for SAT Solving through Substructure Counting S , HS 24 Jeremy Gleixner
Foundation Models for Decoding Brain Activity - Benchmarking B HS 24 Josua Bürki
Multi-Task Learning for Segmentation of Clinical EEG Datasets M HS 24 Jonas Lauer
Symbolic Music Generation with Autoregressive Graph Autoencoder [confidential] S , HS 24 Ziheng Chi
Giving ChatGPT a Virtual Body S HS 24 Han Xi
Human-aligned Compression for Robust Models [confidential] S , HS 24 Samuel Räber
Kantian Moralists and Altruists [confidential] M , Andrei Constantinescu HS 24 Jari Ettlin
Low Fees, High Hopes, Minimal Expectations: Optimistic MEV in the Ethereum Ecosystem [confidential] B , HS 24 Ozan Solmaz
ACORD: An Expert-Annotated Dataset for Contract Clause Retrieval [confidential] M HS 24 Steven Wang
Homo Moralis Plays the Centipede Game [confidential] B , Andrei Constantinescu HS 24 Murezi Giger
Multi-modal Contrastive Learning for Emotion Recognition [confidential] M HS 24 Dong Chu
Physics-Informed Graph Neural Networks for Airflow Simulations [confidential] S , HS 24 Segui Ugalde Aleix
Neural Networks for EEG Data: Compression and Reconstruction Validation [confidential] S , HS 24 Riccardo Rigoni
Self-Supervised Overlap Removal with Graph Neural Networks [confidential] B HS 24 Jan-Erik Gustafsson

Interested in a previous thesis? Show all past theses!

Templates & Guidelines