jnorthrup / photon Goto Github PK

This project is to test the premise of programming in higher dimensions using dimensionally reduced transmission techniques to indicate program state, instructions, instruction effects, object mutation and lifecycle.

For a given program, we typically have a language, parameters, instructions, datatypes, and well defined finite execution properties forming the basis for a thing humans run on computers to perform work.

it is my hope to express these factors in terms of pushing a hypothetical marble through space and declaring a program's state as a collection of trajectories signaled by positive energy changes to a signal, and describe instructions and instruction effects signalled by negative energy along the same dimension-- such as a plane or angular surface(s) that change the properties and manifestation of state along consistent proportional vectors to original imperative code.

A version of opennars software has been used as a research vehicle to measure the fit of a relatively succinct codebase capable of demonstrating learning along limitted dimensional frame of reference. My thesis involves proving learning equivalence in higher dimensional representations and interactions.