Open threads and questions

Experiencing life as an explorerExploration has been a great part of my life, taking exploration here in the most generic / abstract meaning you can give. It is also the main way that I experience interconnection to everything else. It seems to me that we can build very interesting models of ourselves, our senses and the abstractions that make sense by deep introspection and shared communication and experiences, and this seems to be an unmoving goalpost that fills my life with meaning and happiness. Over the years I have been interested in different topics and exploring their interconnection is a way to take advantage of the exploration done, while advancing into new terrains / topics.

I am making this page so that I can have an open / public-facing collection of unanswered question, unfinished projects and on-going research and system building threads.

Models how-to

  • What is the best way incorporate time in our models? How to make it physically realistic to the best of our scientific knowledge
  • What is the best way to incorporate physical constraints into an internal learned simulation of the observation space

The way time is used in probabilistic models and deep learning models can be clustered into the following categories: 1) time as index 2) time as continuous non-random variable 3) time as a random variable. We can incorporate restrictions in any of this to make them realistic (such as positive, or modelling only the deltas). Future direction: time as emerging property of how the other interacting parts of the system evolve dynamically (so here we would have a difference between the dynamical system state evolution time ticking, the time correlated with changes of variables at the observation level maybe)

Humans

  • How can we measure, categorize and start learning the dynamics of internal state loops (from emotions, to thoughts and sensorial information)
  • Can we find structural and repeated patterns of emotional contagion? Which essential dimensions should we look for in order to get a good approximation?

Self-learning

  • Measures of progress
  • Open goals for learning vs explorative goals
  • Inspiration and motivation
  • Pruning, choosing where not to go vs exploring the full space