Minimal Combination for Incremental Evolving Fuzzy Grammar Fragment Learning Nurfadhlina Mohd Sharef, Trevor Martin, Yun Shen Artificial Intelligence Group, University of Bristol, BS8 1TR UK [email protected] , [email protected] , [email protected]
Jun 13, 2015
Minimal Combination for Incremental Evolving Fuzzy
Grammar Fragment LearningNurfadhlina Mohd Sharef, Trevor Martin, Yun Shen
Artificial Intelligence Group, University of Bristol, BS8 1TR [email protected], [email protected], [email protected]
Outline
• Background of problem
• Literature review
• Fuzzy text pattern learning
• Grammar Combination
• Independent Order Grammar Learning
• Conclusion
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
Text Fragments Structure
• Text Fragment Patterns• Similar but not identical• Less Structured• Text segments and simple sentences
• Application: Instance matching, Dialogue system, Question Answering, etc
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
Existing Approaches
• full language model
• tagging-based information extraction
• document distributions and statistical model
• evolutionary genetic algorithms
• Schema/structure matching
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
Grammars for Postal Addressnumber, street name, town, postCode‘21 London Rd Ipswich Suffolk IP1 2EZ’
• And others:• ‘29 Meredith Rd Ipswich’
number, street name,town• ‘Belfairs Hotel 33 Graham Rd Ipswich’
word, business, number, street name, town• mamma's pizza 46 st.matthews st ipswich • beautiful designs 8 norwich rd ipswich
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
Learning Fuzzy Grammar Fragments
“To investigate
an incremental evolving fuzzy grammar fragment learning
with independent-order feature”
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
ISSUES
• Grammar Variety-incremental evolving
• Grammar Derivation-fuzzy grammar, fuzzy membership
• Grammar Similarity- fuzzy grammar similarity
• Grammar combination-maximal, minimal, flexible, compact
• Flexible-independent order
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
Fuzzy Approach for Text Pattern Learning
• Fuzzy Membership• relation between
text and the grammar
fragment
• Fuzzy Grammar
Similarity• Distance
between grammars• <I D S Rs Rt>
ALPHANUMERIC
NUMBER ALPHABETIC
ANYWORD
PC1 PC2
PLACENAME BUSINESS TYPE
ETC
CITY NAME
Figure 5: Partial Order Table for UK Address
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
Grammar SimilaritySource string W E D N E S D A Y
Target string T U E S D A Y
Edit distance* S=1 S=1 D=1 D=1 = = = = =
Table 1: Example of string edit distance operation (*I:Insert, D:Delete, S:Substitute)
Source grammar
Number Word Word Streetending Placename
Target grammar
Number Placename Streetending Placename Countyname
Edit distance*
= S=1 D=1 = = I=1
Table 2: Example of Grammar Edit Distance Operation (*I:Insert, D:Delete, S:Substitute)
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
Start s=new string maxMem=membership(s,TG)
maxMem<1?
Grammar Learning
costST=costTS=1Y
gx=Combine(sg,tg)
costST=1 && costTS=0 gx=sgN
Y
N
costST=0 && costTS=1 gx=tgY
Update TG:GTj={GTj-1-gti} union {gx}
gx=sgN
sg=deriveGrammar(s)tg=target grammar with maxMem
costST=grammarSimilarity(sg,tg)costTS=grammarSimilarity (tg,sg)
End
Y
N
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
Minimal Combination Algorithm
Source Grammar
Target Grammar
Cost(Source,Target)
Cost(Target,Source)
Combination Operation
Combinedgrammar
a-b-c a-b 0 1 0 - - 1 0 0 - - Insert a-b-[c]
a-b a-b-c 1 0 0 - - 0 1 0 - - Insert a-b-[c]
a-b-c a-B-c 0 0 0 - - 0 0 1 - - Merge a-B-c, B>b
a-B-c a-b-c 0 0 1 - - 0 0 0 - - Merge a-B-c, B>b
a-F-c a-G-c 0 0 1 - - 0 0 1 - - Create a-X-c , X:=F||G,GF
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
Experiment-Parsing Coverage• Data
• Restaurant.xml
and yellowpages.xml
• Setting• 25% train, 75% test
• Parameters• # max parsing of
each string• # max parsing by
each grammar
Setting
DatasetAverage
Maximum Parsing (%)
FG Restaurant 89.5
IEG Restaurant 100
FG Yellowpages 97
IEG Yellowpages 100
FG: Fuzzy Grammar
IEG: Incremental Evolving Fuzzy Grammar
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
Experiment-Independent Order Grammar Learning
• To compare parsing coverage
of grammars generated from different orders
G1: a a b cG2: b a cG3: a c G4: a a cG5: a a a
Order 1:
G1-G2-G3-G4-G5
Order 2:
G4-G5-G2-G1-G3
Order 3:
G2-G4-G1-G3-G5
a a [b] c (1+4) a a a (4+5) (b||a) a c (2+4)
[b] a c (2+3) [b] a c (2+3) a a b c (1)
a a a (5) a a b c (1) a c (3)
a a a (5)
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09
Conclusion and Future Work
• Incremental evolving fuzzy grammar fragment learning with independent-order effect
• Latest Achievements• Formal proving of independent order in minimal
combination• Flexible Combination and formal proving• Preliminary Theme Classification
• More Future Works• Focus on softer structures for grammar learning and
theme classification
Minimal Combination for Incremental Evolving
Fuzzy Grammar Fragment Learning, IFSA 09