Daniel Swoboda
Doctoral Researcher in Artificial Intelligence
RWTH Aachen University · Machine Learning & Reasoning
Learning the symbolic structure robots need to plan, reason, and act in a world that won’t sit still.
Recent
- 2026 WorkBenchMark accepted as an oral at the 2026 RoboCup Symposium. →
- 2026 New paper at ICAART 2026 — Making Robots Play by the Rules: The ROS 2 CLIPS-Executive. →
- Jul 2025 RoboCup Logistics League world champions in Salvador, Brazil.
- Since 2024 Co-organising the Aachen Symposium on Representation Learning to Act and Plan. →
About My Research
I'm a PhD student in Computer Science at RWTH Aachen University, advised by Professor Hector Geffner in the Chair of Machine Learning & Reasoning (i6).
My work sits between machine learning and symbolic planning. Lately that means teaching robots to plan and act in cluttered, real-world scenes: learning to guide planning with signals about which actions are physically feasible, and learning to turn abstract symbolic plans into the concrete goals a low-level controller can execute. The throughline is bridging the pattern-recognition strength of deep learning with the transparency and structure of symbolic reasoning.
The long-term goal: AI that is more reliable, transparent, and capable of human-like reasoning — particularly for robots acting in complex, dynamic, real-world environments.
Research areas · Task & Motion Planning · Generalised Planning · Neuro-Symbolic AI · Representation Learning · Robot Manipulation
Selected Publications
- WorkBenchMark: A LEGO-Based Assembly Benchmark with an Assembly-by-Disassembly Baseline for the Smart Manufacturing League RoboCup Symposium · Oral Presentation · 2026
- Making Robots Play by the Rules: The ROS 2 CLIPS-Executive 18th International Conference on Agents and Artificial Intelligence (ICAART) · 2026 · PDF
- From Production Logistics to Smart Manufacturing: The Vision for a New RoboCup Industrial League 2025 RoboCup Symposium, Salvador, Bahia · 2025 · PDF