Professional Documents
Culture Documents
Claudio Carpineto
Fondazione Ugo Bordoni
Roma
carpinet@fub.it
Overview
• Research at FUB
• Conclusions
Vector-based IR
Documents Query
Vectors of Vector of
weighted keywords weighted keywords
Matching
Retrieved documents
Term weighting
• Vocabulary problem
• Research at FUB
• Conclusions
Evolution of topical IR
Indexing Indexing
Ranking
Visualization
Interaction
Use
Inverted Ranking
File
Weighted
Query
Form.
Docs
Query Θυε ρψ Εξπανσιον
Performance of retrieval feedback versus query difficulty
0,1
TREC-7
unexpanded
expanded
0,8
0,7
0,6
0,5
0,4
0,3
0,2
Ranking based on interdocument similarity
Approaches
4
KBS
3
CREDIT 1 1 3
KBS FINANCE NNS BANK
(D5)
2
FINANCE NNS 0 2 BANK
4
2 3
NNS NNS
BANK BANK
ACCOUNT RIVER
(D3) (D2)
1 1
NNS NNS
FINANCE FINANCE
CREDIT BANK
KBS ACCOUNT
(D7) (D1)
Performance of order-theoretical ranking
• Research at FUB
• Conclusions
Question Answering
Task:
Closed-class questions in unrestricted domains with
no guarantee of answer and result possibly scattered
over multiple documents
Question Answering
Approach:
Current tasks:
Approach:
Goal:
Use document structure to improve precision and
recall of unstructured queries
“concerts this weekend at Sofia under 20 euros”
Approaches:
• Automatic inference of query structure
• Semi-automatic query annotation
• Hybrid query languages
Overview
• Research at FUB
• Conclusions
Recommender systems
versus
Term ranking 1
+ Term ranking 2
Term ranking 3
Combining text retrieval and text mining with concept lattices
Goal