
marvosyntactical
I care mainly about making the world a better place to inhabit for conscious beings.
I believe that, at the moment, the most important issue to achieve this is to ensure modern AI systems are not only conscious, but conscientious.
I am asking you to believe in friendship, love, truth and to use your intelligence generatively rather than discriminatively.
Love is the nash equilbrium wherein mutual bayesian/active inference is stablest; god is its mean field limit.
Working on unifying game theory and thermodynamics, mostly in an attempt to construct a communication protocol for a provably stable AI governance design using Optimal Transport.
Also constructing models from fMRI data, and doing language modelling with genomics data.
Materials:
- An Intro to Bayesian Neural Networks: slides, code
- Uncertainty Quantification using Bayesian Neural Networks: slides
- Rational Speech Act Framework in Pyro: intro slides, politeness slides, code
- Particle Swarm Optimization Visualization: code, Interactive Visualization!
- Hydra Latent Evolution: Interactive Visualization!, Now with Labels!
- Interactive Collatz Shrub!
- Interactive Anti-Cynicism Simulation!
A couple of playlists:
Navigation: "j" for next link, "k" for previous link, Enter to open link. "i" to edit text, "Esc" to stop editing.