Special Issue
I am honored to be editor of Springer Astrodynamics' special issue on spAIce 2024: One small step for AI in and for space!
Dominik Dold
Marie Curie Research Fellow
Latest news
Proceedings published
With more than 150 submitted papers and over 200 registrations, SPAICE was a huge success. Check out the proceedings published on Zenodo!
New position
I started as a Marie Curie Research Fellow in the Machine Learning Group of Prof. Philipp Petersen at the University of Vienna!
Joining Dagstuhl seminar
I am very honored to have been invited to an upcoming Dagstuhl seminar on (Actual) Neurosymbolic AI: Combining Deep Learning and Knowledge Graphs!
SpOC 3.0 launched
We just launched the third edition of ESA's Space Optimization Competition! Check out the Programmable Cubes challenge that I designed :)
Invited talk at AI UK 2024
I was invited to give a talk about my work at the Neuromorphic technology: a giant leap for AI workshop of the AI UK Fringe events.
Call for papers: SPAICE
We opened the call for papers of the new conference SPAICE: AI in and for Space. I am super excited to be a chair of the scientific committee!
New journal publication
My publication Differentiable graph-structured models for inverse design of lattice materials has been published in Cell Reports Physical Science.
Marie-Curie Fellowship!
My proposal BASE: Biologically-inspired Autonomous Systems for Space Exploration has been funded under a Marie Curie (MSCA) Fellowship.
Invited talk at UCL AI Society
I had the pleasure to give an invited presentation about my scientific career and my work at the UCL AI Society in front of more than 30 students.
For the previous three years, I have been a Research Fellow at the Advanced Concepts Team of the European Space Agency in Noordwijk. In my work, I investigated neuro-inspired AI and methods of autonomous, functional reconfiguration for a variety of space applications.
Before joining the European Space Agency, I worked as an AI Researcher in Residence at Siemens in Munich, working on graph machine learning for cybersecurity and spike-based graph algorithms.
And even before that, I did a PhD in Physics at Heidelberg University and the University of Bern on neuro-inspired artificial intelligence; mainly focusing on spiking neural networks and neuro-inspired learning rules.
My research interests
Throughout my career, I had the opportunity to dive into many different disciplines such as neuromorphic computing, artificial intelligence (AI), graph representation learning, and materials science - all of which influenced my current research greatly. In general, my research addresses the question how complex systems composed of individual, locally interacting components can optimise global properties of the system as a whole. Or put simply: how does functionality emerge from simple, locally interacting components (and as energy efficient as possible!) - be it biological and artificial neural networks, lattice structures, multi-robot systems, or satellite swarms.
My work can be broadly summarized using the following three categories:
Neuro-inspired AI
I am interested in the interface between modern AI and neuroscience, mostly spiking neural networks, local and decentralised learning rules, continuous learning, and neural development. My aim is to not only better understand information processing in the brain, but to develop novel energy efficient, robust, and adaptive AI methods.
Graph learning
I further focus in my work on graph-structured data and representations and how to process them using neural algorithms, e.g., using graph embedding methods and graph neural networks.
Self-configuring systems
I am enthustiastic about developing novel methods of reorganisation for futuristic space infrastructure such as multi-robot systems, CubeSat swarms, and small-to-large scale lattice structures. This parallels my work on neuro-inspired AI, asking similar questions about local and decentralised functional reorganisation - but for engineered systems instead of biological tissue.