You get a bonus - 1 coin for daily activity. Now you have 1 coin

Restoration of semantic relationships (Causal Reconstruction)

Lecture



Example

l One of several paired, arcuate, flat bones that extend from the spine to the brisket and form the rib cage.

Example

l Edge. One of several paired arcuate flat bones, going from the spine to the chest bone and forming the chest.

l Is there a pair for each edge?

l What is the chest?

l Is the number of edges 2 multiple?

CR (Causal Reconstruction) Task

l Having such a description of an encyclopedia item, the program should be able to answer meaningful questions.

Transition Space

l Consider processes as transition chains

l The perception of causality

l Qualitative changes

l Example

l Conveniently generate from verbal reports

Meaning modeling (Causal modeling)

l Unfiltered input

Tasks of a person doing input to a CR

l Number

l Quality

l Connectivity

l Style

Check created model

Check created model

l Number

l Are new objects added?

l Quality

l Consistency

l Feasibility

l Connectivity

l Creating a semantic relationship on a set of objects

l Style

l Is it possible to “press” the created model?

Simplify 3 types of sentences

l Event

l Defining static connections between objects

l Binding

Additional input

l Add new static links and assertions

l Event definition

l Prior events

l withdrawal rules

l Confirmation Rules

2 levels of understanding of the created model

l Events l Objects

3 types of valid questions. Object level

l Questions related to changing object attributes over time l What happens to the cursor position during Windows shutdown?

3 types of valid questions. Event level

l Relationship between events

l How does the change in the wavelength of light affect the change in the width of the interference strip in the Fraunhofer experiment?

l Possible relationships between the events described in the question and the model already created

l How can a change in humidity affect mood?

Transition space

l Representation of the world by man through events and objects

l Perception of time as a sequence of breakpoints

l Qualitative perception

l Compound change

States vs. Changes (States vs. Changes)

l Example

l Changes are used in speech

Example

l The contact between the steam plate and the metal plate appears .

l The concentration of the solution increases.

l The appearance of the film changes.

l The pin becomes a part of the structure.

l The water remains inside the tank.

Example

l The contact between the steam plate and the metal plate appears .

l The concentration of the solution increases .

l The appearance of the film changes .

l The pin becomes a part of the structure.

l The water remains inside the tank.

Change Classification General

l APPEAR

l DISAPPEAR

l NOT-APPEAR

l NOT-DISAPPEAR

Specialization for NOT-DISAPPEAR

Qualitative attributes

CHANGE NOT-CHANGE

Quantitative Attributes

INCREASE NOT-INCREASE

DECREASE NOT-DECREASE

Predicate notation

l Name of change

l Attribute

l Member Objects

l Time points

Example

l APPEAR (contact, <the-steam, the-metal-plate>, t1, t2)

l INCREASE (concentration, the-solution, t3, t4)

l CHANGE (appearance, the-film, t5, t6)

l APPEAR (a-part-of, <the-pin, the-structure>, t7, t8) l NOT-DISAPPEAR (inside, <the-water, the-tank>, t9, t10)

Grammar. First form

l <input sentence> :: =

<attribute-expression> <verb-group>

l <attribute-expression> :: = the <attribute> <preposition> <noun-phrase> {{<preposition> | and} <noun-phrase>} * l <verb-group> :: = CHANGE | APPEAR, etc. l The concentration of the solution increases.

Grammar. Second form

l <input sentence> :: =

<object> <verb-group> <predicate-modifier>

l <predicate-modifier> :: =

<attribute> [[<<preposition>] <noun-phrase>

{{<preposition> | and} <noun-phrase>} *]

l <verb-group> :: = becomes | becomes not | remains | remains notl The water becomes a vapor.

Links

1. CLEF. http://clef-qa.itc.it/

2. WordNet. http://wordnet.princeton.edu/

3. Pen treebank.

http://www.cis.upenn.edu/~treebank/

4. Start. http://start.csail.mit.edu/

5. TREC. http://trec.nist.gov/

6. Eugene Charniak [1997], “Statistical

Techniques for Natural Language Parsing ”

7. Gary C. Borchardt [1993], “Causal Reconstruction”

Links

8. Boris Katz, Beth Levin [1988] “Exploiting Lexical

Regularities in Designing Natural Language Systems ”

9. Boris Katz and Jimmy Lin. Annotating the Semantic Web Using Natural Language. September, 2002.

10. Boris Katz, Sue Felshin, Deniz Yuret, Ali Ibrahim, Jimmy Lin, Gregory Marton, Alton Jerome McFarland and Baris Temelkuran. Omnibase: Uniform Access to Heterogeneous Data for Question Answering. June, 2002.

11. SEMLP. http://semlp.com/

12. RCO. http://www.rco.ru/

created: 2016-03-11
updated: 2021-03-13
108



Rating 9 of 10. count vote: 2
Are you satisfied?:



Comments


To leave a comment
If you have any suggestion, idea, thanks or comment, feel free to write. We really value feedback and are glad to hear your opinion.
To reply

Creating question and answer systems

Terms: Creating question and answer systems