FrameNet What, how and why

1
FrameNetWhat, how and why

• C. J. Fillmore
• C.F.Baker

2
Who pays us?

• International Computer Science Institute
• National Science Foundation
• Defense Advanced Research Projects Agency

3
Related Projects

• WordNet (Princeton)
• words grouped into synsets
• relations between synsets
• Proposition Bank (Pennsylvania)
• Penn TreeBank
• semantic annotation
• FrameNet
• words related to frames
• valence descriptions uniting frame info and
  syntax

4
Background assumption

• Hypothesis People understand things by
  performing mental operations on what they already
  know. Such knowledge is describable in terms of
  information packets called frames.

5
Lexicon Building

• FrameNet is a lexicon-building project for
  English, relating words to their meanings (via
  the frames that underlie them), recording the
  ways in which phrases and sentences are built
  around them, using the evidence found in large
  samples of modern English usage.

6
The core work of FrameNet

1. characterize frames
2. find words that fit the frames
3. develop descriptive terminology
4. extract sample sentences
5. annotate selected examples
6. derive "valence" descriptions

7
FrameNet Frames

• Intuitively, the frames we mainly work with stand
  for more-or-less fine-grained situation types,
  and the concepts we use in describing them are
  defined relative to the individual frame.
• Here are some examples of frames and their
  constituent frame elements.

8
Finding words that belong in a given frame

• We look for words in the language that bring to
  mind the individual frames.
• We say that the words evoke the frames.

9
Words?

• But first theres an enemy we have to deal with
  polysemy, lexical ambiguity, multiple meanings of
  a single word.
• Instead of words, we work with lexical units
  (LUs), each of these being a pairing of a word
  with a sense.
• In WordNet different LUs belong to different
  synsets in FrameNet different LUs (typically)
  belong to different frames.

10
Polysemy

• FrameNet is at the splitting end of the
  splitting versus lumping continuum when it
  comes to the monosemy/polysemy.
• Its generally assumed that for IR purposes both
  WN and FN make too many distinctions.
• Here are some examples of how we think when
  trying to determine the separateness of lexical
  units with the same form?

11
Discernible meaning differences.

• If a word communicates different meanings
  in different contexts and the difference isnt
  explained by the contexts maybe the word has
  more than one meaning.
• She earns a lot less than she deserves.
• I made a lot of money, but I earned it.
• The second sentence conveys the idea that the
  amount of money earned was appropriate.

12
How many meanings for replace?

• put (sth) back where it belongs
• occupy a position formerly occupied by
  (sth,sbd)
• put something in a position formerly occupied by
  (sth,sbd)

13
John replaced me.
14
John replaced the telephone.
15
Just having different argument types in
grammatical positions isnt enough.

• Subject as Speaker Mom explained , you
  complained , she said , I insist , the dean
  informed us
• Subject as Medium chapter 2 explains , your
  letter complains , the Bible says , the law
  insists , the editorial informed

16
Those dont require separate senses.

• The Medium-as-Subject examples can be thought
  of as Metonymy. Thus
• Chapter 2 explains
• The author explains in Chapter 2 that
• Your letter complains that
• You complain in your letter that

17
Heres a different situation

• Some - but not all - verbs of speaking have a
  cognitive use, identifying sources of beliefs
  or belief-attitudes, with no actual communicating
  implied.
• The heavy winds explain the number of windmills
  around here. (explicate)
• These facts argue in favor of your hypothesis.
  (reason)(quarrel)
• His repeated absence at meetings suggests that
  hes not happy with the job. (hints)

18

• That is, we take the fact that some but not
  all words in a particular semantic class have
  special meaning elaborations argues for a
  polysemy interpretation in those cases.

19
Different Complementation

• Complementation patterns should go with
  particular meanings of a word.
• Medical sense of complainthe patient complained
  of back pains
• Official act sense of complain
• we complained to the manager about X
• she complained that her checks were late

20
Argument omissibility

• We would argue that the ordinary sense of give
  and the contribute sense of give should be
  separated, since they differ in argument
  omissibility
• Do you want to meet the Red Cross representative?
  - I already gave.
• Did you remember a present for your daughters
  birthday? - I already gave.

21
Nominalization differences

• If a verb has two different event noun
  derivatives, and they have different meanings
  that are also found in the verb, the verb itself
  should also be described as polysemous.

22
Nominalization Differences

• adhere to a belief adherenceadhere to your
  skin adhesion
• observe a rule observanceobserve the
  process observation
• commit to a cause commitmentcommit sb to an
  asylum commitmentcommit a crime
  commission
• deliver a package deliverydeliver sb. from
  danger deliverance

23
Support verb differences with nominalizations

• argue quarrel sense associated with have an
  argument reasoning sense with make an argument
• commit dedication sense associated with make a
  commitment crime/sin sense incarceration
  sense, no support verb
• complain symptom report present a complaint
  kvetch no support verb official file a
  complaint, register a complaint, lodge a complaint

24
FN work characterizing frames

• Lets work through the Revenge frame.

25
The Revenge frame

• The Revenge concept involves a situation in
  which
• A has done something to harm B and
• B takes action to harm A in turn
• B's action is carried out independently of any
  legal or other institutional setting

26
Vocabulary for Revenge

• Nouns revenge, vengeance, reprisal, retaliation,
  retribution
• Verbs avenge, revenge, retaliate (against), get
  back (at), get even (with), pay back
• Adjectives vengeful, vindictive
• VN Phrases take revenge, exact retribution,
  wreak vengeance

27
FN work choosing FE names

• We develop a descriptive vocabulary for the
  components of each frame, called frame elements
  (FEs).
• We use FE names in labeling the constituents of
  sentences exhibiting the frame.

28
FEs for Revenge

• Frame Definition Because of some injury to
  something-or-someone important to an avenger
  (maybe himself), the avenger inflicts a
  punishment on the offender. The offender is the
  person responsible for the injury.
• FE List
• avenger,
• offender,
• injury,
• injured_party,
• punishment.

29
Semantic Roles

• Notice that we use such situation-specific
  notions as injury, offender, etc., rather than
  limiting ourselves to some standard list of
  thematic roles, like agent, patient, goal, etc.
• First, there arent enough of those to go around,
  and if we had to squeeze all the distinctions we
  find into such a list,
• we would waste too much time finding criteria to
  do the mapping,
• and we would have to remember what decisions wed
  made.

30
Collecting examples

• We extract from our corpus examples of sentences
  showing the uses of each word in the frame. We
  depend on corpus data rather than existing
  dictionaries or our intuitions about the
  language.
• Our main corpus is the British National Corpus
  we have recently added lots of newswire text from
  the Linguistic Data Consortium. The total is
  about 200M running words.

31
FN work annotating examples

• We select sentences showing all major syntactic
  contexts, giving preference to those with common
  collocations.
• Using the names assigned to FEs in the frame, we
  label the constituents of sentences that express
  these FEs.
• The next slides show what our annotation software
  looks like.

32
The list of frames - including Revenge CLICK
ON Revenge
33
Avenger Offender Injury Injured
Party Punishment
List of FEs CORE Av, InjP, Inj, Off, Pun
List of LUs CLICK ON avenge
34
List of salient contexts for avenge CLICK ON
rcoll-death
35
Sentences with avenge ... death CLICKED
ON sentence 1
36
Annotaters workspace with sentence 1
37
List of FEs for Revenge CLICK ON GF
38
Streamlined list of grammatical functions CLICK
ON PT
39
Two Kinds of Targets

• Predicates
• words that evoke frames, create contexts for
  fillers of information about frame instances
• Fillers
• words that (serve as the heads of constituents
  which) satisfy semantic roles of frames evoked by
  predicates
• many of these evoke frames of their own

40
Separate kinds of annotation

• When targets are predicates
• find the arguments
• When targets are fillers
• find the governor
• find the enclosing phrase
• identify the frame and the FE of that phrase

41
Valence Variation

• We typically find that different words in the
  same frame show variation in how the frame
  elements are grammatically realized.

42
Communication

• Speaker Addressee Topic Message

43
Communication

• Speaker Addressee Topic Message
• We spoke to Harrison about the crisis.

44
Communication

• Speaker Addressee Topic Message
• We informed Harrison of the crisis.

45
Communication

• Speaker Addressee Topic Message
• We told Harrison there was a crisis .

46
Communication

• Speaker Addressee Topic Message
• What did you talk about ?

47
Encoding

• Speaker Message_type Manner

48
Encoding

• Speaker Message_type Manner

She expressed her request rudely.
49
Encoding

• Speaker Message_type Manner

She phrased her answer in this way.
50
Encoding

• Speaker Message_type Manner

How should we word our complaint ?
51
Back to Revenge
52
(No Transcript)
53
(No Transcript)
54
I avenged my brother.
55
I avenged my brothers death.
56
Querying the data ask about the form given the
meaning

• Through various viewers built on the FN database
  we can, for example, ask how particular FEs get
  expressed in sentences evoking a given frame.

57
By what syntactic means is offender realized?

• Sometimes as direct object
• we'll pay you back for that
• Sometimes with the preposition on
• they'll take vengeance on you
• Sometimes with against
• we'll retaliate against them
• Sometimes with with
• she got even with me
• Sometimes with at
• they got back at you

58
By what syntactic means is offender realized?

• Sometimes as direct object
• we'll pay you back for that
• Sometimes with the preposition on
• they'll take vengeance on you
• Sometimes with against
• we'll retaliate against them
• Sometimes with with
• she got even with me
• Sometimes with at
• they got back at you

It's these word-by-word specializations
in FE-marking that make automatic FE
recognition difficult.
59
Querying the data ask about the meaning given
the form

• Or, going from the grammar to the meaning, we
  can choose particular grammatical contexts and
  ask which FEs get expressed in them.

60
What FE is expressed by the object of avenge?

• Sometimes it's the injured_party
• I've got to avenge my brother
• .Sometimes it's the injury
• My life goal is to avenge my brother's murder.

61
Coverage

• Lexical coverage. We want to get all of the
  important words associated with each frame.
• Combinatorics. We want to get all of the
  syntactic patterns in which each word functions
  to express the frame.

62
Frequency data

• We do not ourselves collect frequency data. That
  will wait until methods of automatic tagging get
  perfected.
• In any case, the results will differ according to
  the type of corpus - financial news, children's
  literature, technical manuals, etc.

63
What do we end up with?

• Frame descriptions
• Lexical entries
• Annotations

64
What do we end up with?

• Frame descriptions
• (which some use for situation ontologies)
• Lexical entries
• (which some use for lexicon building)
• Annotations
• (which some use as training corpora for machine
  learning)

65
Outreach

• Other activities
• Success in automating frame analysis of raw text
  would be valuable for IE, MT, NLU various groups
  are experimenting with FN for such purposes.
• Other languages
• There are FrameNets or FrameNet-like projects for
  Spanish, German, Japanese, Swedish, Chinese, and
  apparently Hindi, Romanian, and a few others.