概念依存理論的推理規則

自然語言

前言

簡介

歷史

理論篇

知識表達

語法理論

語意理論

語用理論

方法篇

規則比對

機率統計

神經網路

應用篇

語料建構

全文檢索

自動分類

自動摘要

機器翻譯

問答系統

中文處理

程式篇

交談程式

英漢翻譯

維基語料

搜尋引擎

相關資源

語料辭典

程式工具

相關網站

相關文獻

網頁列表

統計資訊

最新修改

訊息

相關網站

參考文獻

最新修改

簡體版

English

Conceptual Dependency Inference Rule

  • 1. LINGUISTIC-INFERENCE. // slot filling by the default assignment.
    • A given word can be taken to mean that s specific but unmentioned object.
    • Example
      • A buy O from B <=> B sell O to A => money transferred.
      • A drink quite a lot=> alcholic beverage implicit stated.
      • John hit Mary => with hand in implicit.
  • 2. ACT-INFERENCE :
    • Present when an actor and an object occur in a conceptualization without an ACT to connect them, and the object in question has a normal function in the world. In this case the normal function is assumed to be the implicitly referenced ACT.
    • Example
      • John told Mary that he wants a book. (ACT-INFERENCE)=> in order to read it
      • John likes chocolate (ACT-INFERENCE) => in order to eat it
  • 3. TRANS-ENABLE-INFERENCE
    • John told Mary that he wants a book. (ACT-INFERENCE)=> in order to read it (TRANS-ENABLE) => (f) John read the book.
    • John likes chocolate (ACT-INFERENCE)=> in order to eat it (TRANS-ENABLE) => (f) John eat chocolate.
  • 4. RESULT-INFERENCE
    • Whenever a TRANS ACT is present and no information exists that would contradict the inferred result.
    • Example
      • John went to South Dakota. => John was in South Dakota (PTRANS)
      • John told Mary that Bill was a doctor. => Mary knows that Bill was a doctor. (MTRANS)
      • John gave Mary a book. => Mary takes the book. (ATRANS)
  • 5. OBJECT-AFFECT-INFERENCE
    • The physical action that has some affect on the object.
    • Example
      • John hit Mary with a rock. => Rock *PHYSCONT* Mary
  • 6. BELIEF-PATTERN-INFERENCE
    • The given event plus its inferred results fit a belief pattern that has reasons.
    • Example
      • John took an aspirin. (BELIEF) => John was sick.
      • John flattered Mary (BELIEF) => John want Mary do something for him.
  • 7. INSTRUMENTAL-INFERENCE
    • Whenever an ACT has been referenced, its probable instrument can be inferred.
      • INGEST (INSTRUMENTAL) => PTRANS
      • PROPEL (INSTRUMENTAL) => MOVE, GRASP, PROPEL
      • PTRANS (INSTRUMENTAL) => MOVE, PROPEL
      • ATRANS (INSTRUMENTAL) => PTRANS, MTRANS, MOVE
      • CONC (INSTRUMENTAL) => MTRANS
      • MTRANS (INSTRUMENTAL) => MBUILD, SPEAK, ATTAND, MOVE
      • MBUILD (INSTRUMENTAL) => MTRANS
      • EXPEL (INSTRUMENTAL) => MOVE, PROPEL
      • GRASP (INSTRUMENTAL) => MOVE
      • SPEAK (INSTRUMENTAL) => MOVE
      • ATTEND (INSTRUMENTAL) => no , MOVE is often applied.
    • Example
      • John is aware that Fred hit Mary => (aware, CONC) (INSTRUMENTAL) => MTRANS (John got this information (Fred hit Mary) ) => ATTEND eye on it or ATTEND ear on it.
      • John receive the ball. => ATRANS (INSTRUMENTAL) => PTRANS, PROPEL
  • 8. PROPERTY-INFERENCE
    • Whenever an object is intruduced in a sentence, certain subpropositions are being made.
    • Example
      • John hit Mary (PROPERTY) => John and Mary exist, John has hand.
  • 9. SEQUENTIAL-INFERENCE
    • One sentence follows another and they share a subject or a proposition.
    • Example
      • All redheads are obnoxious. & Queen Elizabeth I has red hair. => Queen Elizabeth I is obnoxious.
  • 10. CAUSALITY-INFERENCE
    • Two sentences are connected by an “and” or by their appearing in sequence.
    • Example
      • John hit Mary and she died.
  • 11. BACKWARD-INFERENCE
    • An action has occurred that required another action to precede.
    • Example
      • John ate a banana. => Banana must have been PTRANSed to him at some time.
      • John knows where Mary is => this information mujst have been MTRANSed to John at some time.
  • 12. INTENTION-INFERENCE
    • We assume that a person do something because he want do to it, achieve something.
    • Example
      • John hit Mary (INTENTION, AFFECT) => Mary is hurt pleases John.
      • John told Bill that he wants to go to New York. (INTENTION, RESULT) => being locate in New York will please John., Bill know that.

An integrate example

John hit Mary
(LINGUISTIC) => add “hand” as object of PROPEL
(OBJECT AFFECT) => add causal “receipient (Mary) be hurt”.
(BELIEF PATTERN) => add potential cause of the entire event as Mary DO cause John be hurt cause John be angry.
(INSTRUMENTAL) => add instrument of MOVE “hand”.
(PROPERTY) => add predication that John and Mary exist and that John has hand and that
they were in ther same place at the same time.
(INTENTION) => add that John knwe that it would cause him pleasure if Mary was hurt and that is why he did it.

Four kinds of causal links ,

1. ↑ : causal
2. Result (r) : forward causal rules for physical ACTs.
3. Enable (E): backward causal rules for physical ACTs.
4. Disable (d) : disable
5. Initiate (I): forward causal rules for mental ACTs.
6. Reason (R) : backward causal rules for mental ACTs.
7. <=> dependency

Facebook

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License