If you’d like to contribute your time to the project please sign up on Gitlab.
Scope of the work:
Developers will be using a KIF/CYCL-ish logic language we are developing called PPLLL (Performative Proto Logic Learning Language) and use prolog to finish development and integration of the LogicMOO system.
The Logicmoo system is the developer workspace, and space where the user interacts with AI-controlled robots inside a SWISH MOO and Telnet MUD world in lightly formalized english which goes into a logic representation (KIFF And PPLLL) and can then be presented in many different ways.
More on PLL
PLL is an explanation based language, less declarative and more actively narrative than previous languages such as CycL. We will be better representing the “thoughts and actions behind the words’ rather than the literally the logic atoms as if they are the mechanism itself. PPLLL, just like “performative utterances” not only describe a given reality, but also change the social reality they are describing.
The Project Goals
To build AI, starting with the tools and infrastructure a logical being would need to occupy and model its own world to maintain its own codebase
(The main goal of the AI system will be to classify its information & action streams, to act as a perception rationalizer similar to a human it will need several internal dialogues taking place, all of which will be represented in PLL.)
The AI itself resides within a framework that is both Narrative and Animatory. Such that it will seek to rationalize “this is the animation I am currently involved in”
This animation is based on language that evokes an imagined situation (which can be represented in the MOO) and how it can predict this framework to play out. This is done explicitly NOT at the level of the egg cracking problem, but more conceptually and therefore parsimoniously. The AI will be constantly reminding & guiding itself as to what it’s currently doing. In the background it will add and explore neighboring/elaborating details about the current animation, analogous to forward chaining, but constrained by effort & appropriateness.
The system has several very similar 4-Layer Components share much of the same code but have different functions they perform in the system: