What are the different ways that a molecule can be encoded into a vector? My answer to that question would employ machine learning but maybe there are other approaches, perhaps analytical, e.g., graph theoretical. Are there specific examples in scientific literature?

Would it be possible to use a molecular vectorization in many-body quantum mechanics (effectively computational quantum mechanics)? A potential example that I can offer is to use a molecular vectorization to create a mapping to the eigenvectors in the eigenvalue problem in DFT.

  • 2
    $\begingroup$ What properties of the molecule would you like to describe with a vector? What are your criteria for that vector? Have you checked any example in the literature? $\endgroup$
    – Greg
    Jun 9 at 0:49
  • $\begingroup$ I'm not quite sure yet about which properties to consider. I just want to come up with some sort of blueprint that leads me to the eigenvectors in the (DFT) Schrodinger Equation. At the very least, obtain computational speedups. Machine learning can obtain satisfactory results in approximating functions; I was thinking about doing something more targeted to start off. To ground my thinking, I'll offer node2vec and graph2vec as examples. $\endgroup$
    – Jonathan
    Jun 11 at 21:16
  • 2
    $\begingroup$ This is a very active field of research, and you can find reviews, tutorials, workshops. I recommend you to start with a specific problem $\endgroup$
    – Greg
    Jun 12 at 5:51
  • $\begingroup$ @Jonathan have you been able to narrow down your question to a more specific problem, so that perhaps Greg can write an answer? $\endgroup$ Jul 12 at 6:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.