text meaning

Views:
 
Category: Entertainment
     
 

Presentation Description

No description available.

Comments

Presentation Transcript

A Knowledge-Driven Approach to Meaning Processing: 

A Knowledge-Driven Approach to Meaning Processing Peter Clark Phil Harrison John Thompson Boeing Mathematics and Computing Technology

Goal: 

Goal Answer questions about text, especially questions which go beyond explicitly stated facts Suggests: there a rocket launch China owns the satellite the satellite is for monitoring weather the orbit is around the Earth etc. None of these are explicitly stated in the text “China launched a meteorological satellite into orbit Wednesday, the first of five weather guardians to be sent into the skies before 2008.”

Meaning Processing: 

Meaning Processing May include elements not explicit in the text Operational definition: Degree of “captured meaning”  ability to answer questions about the scenario being described Meaning processing  construction of a situation-specific representation of the scenario described by text.

Underlying Philosophy: 

Underlying Philosophy Meaning processing is fundamentally a modeling activity: matching what is said with our expectations about the way the world can be creating (assembling) a model of the scenario being described from a library of pre-built, prototypical scenarios Somewhat of a “Schankian” approach…

Approach: 

Approach Heavily expectation-based: Build knowledge-base of prototypical scenarios e.g., “launching a satellite” Extract fragments of information from text e.g. subject-verb-object tuples Find scenario(s) best matching the text drives disambiguation and coersion of the text scenario provides info unstated in the text

Scenario Representations: 

Scenario Representations Identify and label the “key participants” Encode relationships between them in a KR language Use of prototypes Using WordNet’s “ontology” (pragmatic decision)

Scenario Representation (simplified) (encoded in the KM language): 

Scenario Representation (simplified) (encoded in the KM language) (launch-a-satellite_v1 has (superclasses (launch_v1 transport_v1)) (text-definition ("An event that begins when ….”))) (every launch-a-satellite_v1 has (step_n1 ((a countdown_n1 with (location_n1 ((the location_n1 of Self))) (event_n1 ((the fly_v1 step_n1 of Self))) (before_r1 ((the fly_v1 step_n1 of Self)))) (a fly_v1 with (vehicle_n1 ((the vehicle_n1 of Self))) ))) (vehicle_n1 ((a rocket_n1))) (cargo_n1 ((a satellite_n1))) (locaton_n1 ((a launchpad_n1))) (agent_n1 ((a entity_n1)))))

Slide8: 

place_v1 entity_n1 agent_n1 entity_n1 object_n3 destination_n1 step_n1 place_n1 destination_n1 move_v1 release_v1 before_r1 object_n3 + Representations are Compositional (1) location_n1 launch-a-satellite_v1 countdown_n1 rocket_n1 step_n1 orbit_n1 destination_n1 fly_v1 entity_n1 vehicle_n1 agent_n1 launchpad_n1 location_n1 before_r1 vehicle_n1 agent_n1 agent_n1 satellite_n1 cargo_n1 cargo_n1 release_v1 before_r1 destination_n1 location_n1 launch-a-satellite_v1 countdown_n1 rocket_n1 step_n1 orbit_n1 destination_n1 fly_v1 entity_n1 vehicle_n1 agent_n1 launchpad_n1 location_n1 before_r1 vehicle_n1 satellite_n1 cargo_n1 cargo_n1

Representations are Compositional (2): 

Representations are Compositional (2) Create multiple representations for a single concept each encoding a different aspect/viewpoint Representations can be combined as needed Objects Involved Temporal location_n1 launch-a-satellite_v1 countdown_n1 rocket_n1 step_n1 orbit_n1 destination_n1 fly_v1 entity_n1 vehicle_n1 agent_n1 launchpad_n1 location_n1 before_r1 vehicle_n1 agent_n1 agent_n1 satellite_n1 cargo_n1 cargo_n1 launch-a-satellite_v1 countdown_n1 rocket_n1 step_n1 fly_v1 vehicle_n1 launchpad_n1 location_n1 vehicle_n1 agent_n1 satellite_n1 cargo_n1 launch-a-satellite_v1 countdown_n1 step_n1 before_r1 vehicle_n1 fly_v1

Scenario Representations: 

Scenario Representations Goal is to capture “intermediate level” scenes: more specific than “Move” or “Launch” more general than “The launching of meteorological satellites by China from Taiyuan.” Setting-In-Motion Launching-A-Satellite China-Launching-A-Weather-Satellite The Intermediate Level

Approach: 

Approach Heavily expectation-based: Build knowledge-base of prototypical scenarios e.g., “launching a satellite” Extract fragments of information from text e.g. subject-verb-object tuples Find scenario(s) best matching the text drives disambiguation and coersion of the text scenario provides info unstated in the text

Evidence Extraction from Text: 

Evidence Extraction from Text Use SAPIR (bottom-up chart parser) “Tuples” extracted from parse tree easy to manipulate much semantic interpretation deferred word senses, semantic relations (S NIL “launch” “ship”) (S “ship” “carry” “equipment”) (S “ship” “carry” “food”) (S “ship” “carry” “fuel”) (S “ship” “enter” “orbit”) “A Russian cargo ship carrying equipment, food, and fuel for the Space Station was launched from the cosmodrome on Monday. The blastoff took place at 11am, and the cargo ship entered its orbit 10 minutes later.”

Evidence Extraction from Text (cont): 

Evidence Extraction from Text (cont) “A Russian cargo ship carrying equipment, food, and fuel for the Space Station was launched from the cosmodrome on Monday. The blastoff took place at 11am, and the cargo ship entered its orbit 10 minutes later.” (NA "ship" "russian") (S "ship" "carry" “equipment" ("for" "station")) (S "ship" "carry" “food" ("for" "station")) (S "ship" "carry" “fuel" ("for" "station")) (S NIL "launch" "ship" ("from" "cosmodrome") ("on" "monday"))) (S "ship" "enter" "orbit") (S "blastoff" "take place" ("at" "id2")) (more recent tuple extractor)

Approach: 

Approach Heavily expectation-based: Build knowledge-base of prototypical scenarios e.g., “launching a satellite” Extract fragments of information from text e.g. subject-verb-object tuples Find scenario(s) best matching the text drives disambiguation and coersion of the text scenario provides info unstated in the text

Matching the Text with Scenarios: 

Matching the Text with Scenarios A syntactic tuple matches a semantic assertion if: its words have an interpretation matching the concepts the syntactic relation can map to the semantic relation Simple scoring function f(# word matches, # tuple matches) (S “china” “launch” “satellite”) “china” “launch” “satellite” “launch” subject direct object location_n1 launch-a-satellite_v1 countdown_n1 rocket_n1 step_n1 orbit_n1 destination_n1 fly_v1 causal_agent_n1 vehicle_n1 agent_n1 launchpad_n1 location_n1 before_r1 vehicle_n1 agent_n1 agent_n1 satellite_n1 cargo_n1 cargo_n1

Multiple abstractions may match…: 

Placing-In-Position Launching-A-Satellite Moving Orbit-Celestial Rocket-Vehicle Artificial-Satellite Tangible-Entity Launching-A-Boat Launching Boat Multiple abstractions may match… China’s Launch of the FY-1D meteorological satellite

Illustration: 

“A Russian cargo ship carrying equipment, food, and fuel for the Space Station was launched from the cosmodrome on Monday. The blastoff took place at 11am, and the cargo ship entered its orbit 10 minutes later.” Illustration

Illustration: 

“A Russian cargo ship carrying equipment, food, and fuel for the Space Station was launched from the cosmodrome on Monday. The blastoff took place at 11am, and the cargo ship entered its orbit 10 minutes later.” (NA "ship" "russian") (S "ship" "carry" “equipment" ("for" "station")) (S "ship" "carry" “food" ("for" "station")) (S "ship" "carry" “fuel" ("for" "station")) (S NIL "launch" "ship" ("from" "cosmodrome") ("on" "monday"))) (S "ship" "enter" "orbit") (S "blastoff" "take place" ("at" "id2")) location_n1 launch-a-rocket_v1 countdown_n1 rocket_n1 step_n1 orbit_n1 destination_n1 fly_v1 entity_n1 vehicle_n1 agent_n1 launchpad_n1 location_n1 before_r1 vehicle_n1 agent_n1 agent_n1 release_v1 before_r1 destination_n1 Score: 24 Illustration

Issues : 

Issues Constraints to overcome: Language: Requires just the right text Knowledge: Requires just the right representation enter sit order “John entered the restaurant. John sat down. He ordered some food…” “John picked up the menu…”

1. Language: Matching with scenarios: 

1. Language: Matching with scenarios “Ideal”: semantic structure mirrors syntactic structure each syntactic unit has a direct semantic correlate (“china” “launch” “satellite”) entity launch satellite More usual case: no correlate (no direct match) can enumerate types of mismatch can enumerate “rewrite rules” to handle them cargo agent

5 Classes of Mismatch: 

5 Classes of Mismatch Substitutability (synonyms, hypernyms) “the satellite was put in orbit” “the satellite was placed in orbit” Nominalization “… was launched with approval from the regulator”  “the regulator approved the launch” Roles “The State Department was the regulator…”  “The State Department regulated…” Expansions/contractions “supplies fuel to”  “fuels”, “enters orbit”  “orbits” Metonymy “Bush authorized tax cuts…”  “the White House authorized tax cuts…”

2. Knowledge: What’s in a scenario?: 

2. Knowledge: What’s in a scenario? There is no single scenario representation Rather, scenarios have multiple facets Often, multiple scenarios are active at once “Amid great expectations, the anti-impotence drug Viagra was launched in China on Monday with approvals from the State drug regulator. At a press conference, Pfizer Inc. announced China had approved the sale of the drug in the country.”

Slide23: 

2. Knowledge: What’s in a scenario? Implies: Need compositional representations detailed scenario needs to be build from multiple, abstract ones hence need a mechanism for triggering and integrating them Need a lot of knowledge Too time-consuming to build by hand! much of it may be “mundane” Possible sources: WordNet (concept-word mappings, hypernyms) Extraction from corpora (Schubert-style) Dictionary definitions

2. Knowledge: How to build the KB?: 

2. Knowledge: How to build the KB? Many of these facts are “mundane” Can we extract these facts semi-automatically? yes! Can we use them to rapidly assemble models? perhaps ? “satellites can be launched” “rockets can carry satellites” “countdowns can be at launchpads” “rockets can fly”

Semi-automatic extraction of facts…: 

Applying a Schubert-inspired method to the Reuters corpus, we get (for example)… Semi-automatic extraction of facts…

Semi-automatic extraction of facts…: 

Semi-automatic extraction of facts…

Summary: 

Summary Text understanding = a modeling activity Text suggests scenario models to use Models suggest ways of interpreting text Approach: library of pre-built scenario representations NL-processed fragments of input text Matching process to find best-matching scenario Issues (many!); in particular: matching syntax and semantics constructing the KB

authorStream Live Help