The year is coming to an end, Christmas is around the corner, and reviews of 2017’s events are popping up everywhere. I think this is a nice opportunity to also look back at the year 2017, to summarize what has happened in my academic life, and to speculate a bit about 2018.
So what has happened this year? Quite a lot, as it turns out:
- First of all, I created my web presence: This website, an ORCID, and profiles on Twitter, GoogleScholar, ResearchGate, etc.
- I developed a mathematical formalization of the conceptual spaces framework for AI purposes, including a publicly available implementation.
- I managed to publish four peer-reviewed papers (two workshop and two conference papers) about my ongoing research.
- I did quite some traveling and attended multiple scientific events which allowed me to present my own work, to learn about exciting research done by others, and to start growing a network:
- The spring school “Interdisciplinary College” in Günne/Germany
- The Dagstuhl seminar “Human-Like Neural-Symbolic Computation” in Dagstuhl/Germany
- The workshop “Neural-Symbolic Learning and Reasoning” in London/UK
- The “German Conference on Artificial Intelligence” in Dortmund/Germany
- The workshop “Concept Learning and Reasoning in Conceptual Spaces” in Bochum/Germany
- The workshop “Artificial Intelligence and Cognition” in Larnaca/Cyprus
- The “SGAI International Conference on Artificial Intelligence” in Cambridge/UK.
- Together with some colleagues of mine here in Osnabrück I planned an interdisciplinary summer school on concepts for 2018 (with a co-located workshop on the same topic). Currently, we are waiting for the feedback on our funding application.
- I started getting my feet wet with (deep) neural networks: On the one hand with InfoGAN for learning interpretable shape dimensions, on the other hand with Logic Tensor Networks for learning concepts.
- I also started taking didactics classes which have already helped me a lot to improve my teaching.
So what will the year 2018 look like? I’ve given up making detailed schedules because research tasks usually take much longer than I anticipate anyways. But here are some rough ideas:
- I want to continue playing around with InfoGAN, β-VAE, and other representation learning approaches in order to learn meaningful dimensions for describing shapes.
- Moreover, I’ll invest more time into Logic Tensor Networks – adapting them in such a way that they work with my formalization of conceptual spaces and doing some actual concept learning experiments.
- Teaching will also continue to play an important role in my academic life. I’m looking forward to applying the knowledge I’ve gained in the didactics classes in practice.
- There are a few open ends with respect to my formalization that I would like to tie up – namely how to generalize the definitions of similarity and betweenness from instances to concepts.
- After (hopefully) having made satisfactory progress with InfoGAN and LTN, I want to start with the main part of my PhD project: Devising an incremental clustering algorithm for concept formation in conceptual spaces.
I don’t know how realistic these plans are, so we’ll see how many of them I can cross of from this list a year from now.