Achref Jaziri

Center for Cognition and Computation, Goethe University Frankfurt

ProfilePic1.jpg

Room S2|210

Robert-Mayer-Str. 10–12

Frankfurt am Main, Germany

I am a PhD candidate in the Center for Cognition and Computation at Goethe University Frankfurt and a Student Researcher at Google Research. My research focuses on the design of robust artificial intelligence (AI) systems based on principles of systems engineering. During my doctoral studies, I investigate topics at the intersection of causality, continual learning, and the development of robust online systems. I am particularly interested in applying continually evolving systems to real-world problems within the scope of the KIBA project. Prior to my PhD, I completed a Master’s degree in Computer Science with a specialization in Theoretical Neuroscience, working on computer vision models and simulation tools for analyzing concrete defects as part of the EU H2020 RESIST project (RE-Silient transport InfraSTructure to extreme events).

news

Sep, 2025 The KIBA project has concluded, culminating in a successful prototype for forecasting and freight transport planning.
Aug, 2025 Two papers accepted at CoLLAs 2025 — Conference Track: “Mitigating the Stability–Plasticity Dilemma in Adaptive Train Scheduling with Curriculum-Driven Continual DQN Expansion”; Workshop Track: “A Simple Baseline for Stable and Plastic Neural Networks”.
Jul, 2025 Joined Google Research (Student Researcher).
Jun, 2025 synth-dacl: Does Synthetic Defect Data Enhance Segmentation Accuracy and Robustness for Real-World Bridge Inspections? accepted at DAGM GCPR 2025.
Sep, 2024 Attending ECCV 2024.
Aug, 2024 ‘Representation Learning in a Decomposed Encoder Design for Bio-inspired Hebbian Learning’ accepted to the ECCV 2024 Human Inspired Vision Workshop.
Aug, 2024 Preprint for “Mitigating the Stability-Plasticity Dilemma in Adaptive Train Scheduling with Curriculum-Driven Continual DQN Expansion” is out on arXiv.
Jan, 2024 Attending WACV 2024.
Dec, 2023 “Lightweight Techniques to Improve Generalization and Robustness of U-Net Based Networks for Pulmonary Lobe Segmentation” accepted in Bioengineering.
Nov, 2023 Preprint for “Representation Learning in a Decomposed Encoder Design for Bio-inspired Hebbian Learning” is out on arXiv.
Oct, 2023 “Designing a Hybrid Neural System to Learn Real-world Crack Segmentation from Fractal-based Simulation” accepted to WACV 2024.
Jan, 2023 Officially joined the KIBA project.

Publications

  1. synth-dacl: Does Synthetic Defect Data Enhance Segmentation Accuracy and Robustness for Real-World Bridge Inspections?
    Johannes Flotzinger, Fabian Deuser, Achref Jaziri, Heiko Neumann, Norbert Oswald, Visvanathan Ramesh, Thomas Braml DAGM German Conference on Pattern Recognition (GCPR), 2025
  2. Mitigating the Stability–Plasticity Dilemma in Adaptive Train Scheduling with Curriculum-Driven Continual DQN Expansion
    Achref Jaziri, Etienne Künzel, Visvanathan Ramesh Conference on Lifelong Learning Agents (CoLLAs), 2025
  3. A Simple Baseline for Stable and Plastic Neural Networks
    Étienne Künzel, Achref Jaziri, Visvanathan Ramesh CoLLAs 2025 – Workshop Track, 2025
  4. Representation Learning in a Decomposed Encoder Design for Bio-Inspired Hebbian Learning
    A. Jaziri, S. Ditzel, I. Pliushch, Visvanathan Ramesh ECCV 2024 Workshops (Human-Inspired Vision), 2024
  5. Designing a Hybrid Neural System to Learn Real-world Crack Segmentation from Fractal-based Simulation
    A. Jaziri, M. Mundt, A. Rodriguez, Visvanathan Ramesh IEEE/CVF WACV, 2024
  6. Lightweight Techniques to Improve Generalization and Robustness of U-Net-Based Networks for Pulmonary Lobe Segmentation
    Armin A. Dadras, Achref Jaziri, Eric Frodl, Thomas J. Vogl, Julia Dietz, Andreas M. Bucher Bioengineering, 2023