Skip to content

Latest commit

 

History

History
53 lines (36 loc) · 1.47 KB

TODO.md

File metadata and controls

53 lines (36 loc) · 1.47 KB

Do not forget to scale back the scope ~

  • build a causal inference engine for real valued random variables first.

  • try to model a system of random variables

    • review docs/articles/random-variables.md
    • a Sample class?

  • A good API for working with (directed) graph.

  • Thinking about structural causal model, what if the functions are symbolic instead of numerical, we can use pattern matching to do queries, just like we can use inequality to do queries for numerical function.

  • Survey classical expert systems.

    If we have

    Man(x) -> Mortal(x)

    and we ask why Mortal(Socrates), the robot answers because Man(Socrates), the word "why" is used here, but it is not about causality? it is only about explanation, but is explanation about causality?

    if it is our model, the model is not about random variables, but about properties of objects.

    one object might has different properties, and the properties might related by functions.

    note that we can view isMan and isMortal as properties of Socrates of type Boolean.

learn