LISP has been the lingua franca of symbolic AI. Symbolic AI is a paradigm to create algorithms which solve problems by interacting with a symbolic representation of the world.
LISP was a natural choice due to use in teaching and Meta programming facilitates such as the ability to introspect code as data and create new generated code.
Julia's design is inspired by LISP. However Julia poses new challenges. This talk explores whether they can be overcome and if Julia is suitable for Symbolic AI.
Julia strikes an interesting balance. It is designed to be compiled into fast code by limiting dynamism at run time while still allowing REPL based development. Thus it is not obvious if Julia is suitable for symbolic AI.
Julia has a world age mechanism which makes it more difficult to call dynamically defined functions, and it's internal representation of code is more complicated than LISP's, which might present further challenges.
In this talk is based on experience gained while implementing a toy symbolic AI. The presenter will discuss whether more complex and self modifying symbolic AI would be possible to achieve in Julia. If this is the case there should be no problem implementing other symbolic AIs in Julia.