opennars / Narjure

A Clojure implementation of the Non-Axiomatic Reasoning System proposed by Pei Wang.
GNU General Public License v2.0
44 stars 11 forks source link

Remaining TODO #42

Closed patham9 closed 8 years ago

patham9 commented 8 years ago

Besides having a first working reasoner since today already, we are approaching the end of the development process. This issue shows the remaining things to do, from which I expect to be finished in the next two weeks with the help of Lense:

  1. Termlink based selection for general inference. (this I will finish next)
  2. Concept: Local Inference (with decision making rule, casting to operator executor) (Page 11 of design document)
  3. Make sure budget functions are implemented and applied as shown in the pdf. (Page 15 of design document)
  4. General-inferencer: Besides making sure that evidence trail is kept track of, don't even attempt derivation if there is evidental overlap between task and belief.
  5. Deriver: Apply project-eternalize on belief dependent on whether the rule has measure-time-predicate or not. (this I will do)
  6. task creator: Also create a eternalized task for each no matter from where incoming (derived or not derived) sentence message.
  7. GUI: Define a output block in lense, with corresponding "Clojure to Narsese string" translation which can easily be recursively defined, and which display atom is used by task-message.
patham9 commented 8 years ago
    1. was done by Tony today already.
patham9 commented 8 years ago

TODO left:

  1. Termlink based selection for general inference. This I will do next but we could discuss it in the IRC.
  2. Question/Quest answer detection. (both, the concept-local (use as best solution if higher confidence than existing solution) as well as concept-interacting case (use as best solution if higher function value where function is dependent on complexity and truth expectation, see paper) for question-variable-question-answering). The question-var answering I will do.
  3. Make sure budget functions are implemented (also increase-budget decrease-budget in the local inference files) and applied as shown in the pdf. (Page 15 of design document) <- Most is already done (forgetting). Derived task budget and the for it needed termlink associativity I will do.
patham9 commented 8 years ago

these points were all implemed weeks ago already and improved since then.