- His submitted thesis is entitled From Continua to Objecthood: On Tokenization and Perceptual Grouping in Vision.
- By allowing AI models to learn how images can be divided into meaningful regions, they can learn to prioritize structure, shape, and relationships in a way that is closer to human vision. The result of this research is new vision systems that can be more interpretable, efficient, and flexible.
- He will also give the trial lecture entitled From Convolutions to Transformers: How Modern Vision Models See Images
Official announcement with time and location for both trial lecture and defence: Department web news item