Attending my first conference

In December, I attended the NeurIPS 2021 Conference and had the amazing opportunity to present my paper on Compositional Modeling of Nonlinear Dynamical Systems with ODE-based Random Features. Submitting this work in the summer of 2021 was the culmination of over a year of work which began whilst working on my MSc dissertation project in the department, so it was extremely satisfying to have the paper accepted and see all of that hard work come to fruition.

This was my first experience of having a discussion with reviewers, and as the work sits at the intersection of three quite broad and distinct research areas in physics-informed machine learning, probabilistic machine learning and deep learning, I wasn’t quite sure what to expect during the review phase. Additionally, given that this was the first research project I worked on during my PhD, I really had no concept of how the paper would be received. Doing research is quite a gruelling process at times with lots of trial and error, and whilst having great supervisors helps make the journey considerably smoother, getting positive comments from other experts in the field in the form of reviews was such a confidence boost for me as a 1st year PhD student. Overall, the feedback I received was constructive and really quite insightful, and I’ve definitely had a bunch of ideas off the back of my discussions with the reviewers.

After the paper was accepted, given the virtual nature of the conference, I had to put together a video presentation explaining the key ideas in the paper, as well as a poster to present during the conference. Having started my PhD during the pandemic, I’ve had limited opportunities to interact and socialise with other members of the research community, therefore I had been hoping to be able to attend NeurIPS in-person for this reason. Fortunately however, I did still have the chance to chat with some other researchers about my work during my virtual poster session, which was great fun.

At the start of the paper writing process the prospect of having a paper accepted at NeurIPS seemed quite daunting and unattainable, but as I kept working away at the paper and moving through the submission process, I felt more and more confident in my work. Based on my experience, my advice to students looking to submit to a conference would be to not second guess yourself, as I found it can be quite easy to fall into that imposter syndrome trap, especially if it’s your first time. As I mentioned earlier, I am fortunate to have great supervisors in Mauricio Álvarez and Mike Smith who provided me with so much guidance at every stage of the project.

The link to the paper is available here, and for more information about me and my research check out my website here.