Concrete examples of OpenCog's functionality

Question

Does anyone know what specific tasks the OpenCog environment is capable of performing? I have glanced though their wiki and a few of the pages on Goertzel's site and the AI.SE. So far I could only find some technical documentation regarding theory and engineering, but nothing on concrete results.

From the technical description of AtomSpaces it seems that OpenCog is capable of some "representational inference", but I haven't come across any sources that concretely describes what it is capable of doing.

Apparently there is some collaboration between Sophia the Robot and OpenCog, but to what extent I am unclear. I am aware however that the dialogue functions is powered by ChatScript (though I also suspect that the high profile interviews Sophia gives are completely scripted...)

Can anyone provide concrete examples or evidence of OpenCogs' functional behavior. Like transcripts of chat, examples of reasoning, video or demonstrations of its emotion-emulating; and not just claims of functions.

Answer 1

It depends what you mean by "what OpenCog can do?".

OpenCog, at a high-level, is a loosely coupled collection of various theoretical and variants of conventional methods aimed at constructing the beginnings of an AGI.

With that said, it's purely applied uses are fairly limited. It can do some typical NLP and ML tasks if used correctly, albeit, it will almost always be ineffective when compared to a problem specific solution. However, this is understandable, as OpenCog is not geared at solving narrow problems.

OpenCog, as well as Goertzel's other projects(SingularityNet, AGI society), are broad, top-down attempts at formulating an AGI system. Outside of AGI circles they are also not oft referenced or researched(although, I am utilizing some aspects of the system in work I am currently crafting).