While the performance of multimodal world models has significantly improved, they still lack commonsense abstraction and reasoning capabilities. These models fail to capture the causal aspects of the physical and social world. To address the grand challenge of commonsense abstraction and reasoning with multimodal world models, methods will be explored to enhance their robustness, abstraction, and visual grounding abilities.
The focus will be on learning explicit, effective abstractions from limited human-labeled data, serving as dynamic, neurosymbolic world models. Following cognitive principles, these methods will be evaluated on tasks that require commonsense reasoning over space, time, and causality. This project is funded by a talent project and will be co-supervised by researchers from an academic institution. The Postdoc will be part of a European AI research network.
What you will do.
- Show independence in achieving research goals and willingness to collaborate and supervise PhD students working on multimodal abstraction, commonsense reasoning, and world models.
- Invent, evaluate, and describe novel techniques for multimodal abstraction and explicit reasoning as dynamic, neurosymbolic world models.
- Publish and present this research at international conferences, workshops, and journals.
- Become an active member of the research community and collaborate with other researchers, both nationally and internationally.
- Contribute to teaching activities, such as lectures, lab courses, or supervising bachelor and master students.
- Develop your portfolio and vision towards your next career position in academia.
Profile
- You have a PhD in Machine Learning, Computer Vision, Natural Language Processing, Neurosymbolic AI, or a closely related area.
- You can invent and evaluate novel algorithms and present these orally and in writing.
- You have a record of publishing in relevant, high-quality conferences or journals in the above fields.
- You are eager to tackle fundamental challenges in the areas of multimodality, abstraction, commonsense reasoning, and neurosymbolic world models.
- You can implement and evaluate commonsense reasoning methods with multimodal foundation models, e.g., using Python deep learning toolkits.
- You can work well in teams and communicate effectively in written and spoken English.
How to apply
View the full assignment text and application details once your tailored application is ready.
Order a tailored application to view the full assignment and application details.
More context, less searching.
You get enough context to judge whether this job is relevant. The full brief, client details and next steps stay available inside the app.