* Learning and building a more intuitive understanding of Projective Geometric Algebra (PGA). Looks very disruptive for all kinds of computer graphics applications. PGA replaces the use of Vectors, Quaternions, Dual Quaternions, and the entire machinery of Linear Algebra with a single unified, elegant framework that "just works." Feels a bit like magic.
* Exploring and playing with new capsule routing algorithms in deep learning models for vision and language tasks. Particularly intrigued by routing algorithms in which output capsules seek to "explain" (generate/predict) their input data (e.g., EM matrix routing, Heinsen routing).
The following video has a good, developer-friendly introduction to the subject that should answer your question. Take a look in particular at the code snippets that are sprinkled throughout the video and notice how much cleaner, shorter, simpler, less computationally costly, and "universal"/"exception-free" the code becomes: https://www.youtube.com/watch?v=tX4H_ctggYo
* Exploring and playing with new capsule routing algorithms in deep learning models for vision and language tasks. Particularly intrigued by routing algorithms in which output capsules seek to "explain" (generate/predict) their input data (e.g., EM matrix routing, Heinsen routing).