Learn publicly. Sounds intimidating, but is likely to be one of the most efficient ways to start getting things done. It is easier and more sometimes more efficient to get feedback than to ask open-ended questions.

I haven't made a lot of progress on many of my research ideas over the past five years. Many of the concepts that excited me back then are still relatively unexplored. Every year that I wait it becomes easier though, as more of the underlying concepts become more polished, and the challenge becomes more of application and execution rather than novel algorithmic research and implementation.

Between today and the end of May 2020, I plan to study the following two papers:

Generalised Measures of Multivariate Information Content
The entropy of a pair of random variables is commonly depicted using a Venn diagram. This representation is potentially misleading, however, since the multivariate mutual information can be negative. This paper presents new measures of multivariate information content that can be accurately depicted…
CausalVAE: Structured Causal Disentanglement in Variational Autoencoder
Learning disentanglement aims at finding a low dimensional representation, which consists of multiple explanatory and generative factors of the observational data. The framework of variational autoencoder is commonly used to disentangle independent factors from observations. However, in real scenari…

There will be many threads to pull from here on, but I plan to focus on a pragmatic experimentation approach for applying these concepts to the chemical process monitoring and control.

I am also excited about the possibilities of generative design and intuitive AI, and plan to start experimenting how one could interface some of the basic algorithms with a mature simulation environment such as Elmer.