UA in ImageCLEF 2005 Maximiliano Saiz Noeda. Index System  Indexing  Retrieval Image category classification  Building  Use Experiments and results.

Slides:



Advertisements
Similar presentations
Multilinguality & Semantic Search Eelco Mossel (University of Hamburg) Review Meeting, January 2008, Zürich.
Advertisements

Complex queries in the PATENTSCOPE search system Cyberspace September 2013 Sandrine Ammann Marketing & Communications Officer.
SINAI-GIR A Multilingual Geographical IR System University of Jaén (Spain) José Manuel Perea Ortega CLEF 2008, 18 September, Aarhus (Denmark) Computer.
Chapter 5: Introduction to Information Retrieval
What are the Spanish School Languages?. STATE LANGUAGE SCHOOLS LANGUAGES TAUGHT - All the official ones in Spain. - Fourteen languages of the European.
© ISIS How do we talk to zoo keepers? How do we talk to zoo keepers? Michele Peters University of St. Thomas.
Sheffield at ImageCLEF 2003 Paul Clough and Mark Sanderson.
Multilingual experiments of CLEF 2003 Eija Airio, Heikki Keskustalo, Turid Hedlund, Ari Pirkola University of Tampere, Finland Department of Information.
Presentation Title Presentation Subtitle and/or Conference Name Place Day Month Year First Name Last Name Job Title.
1 Towards a better transcultural understanding through distance education. Towards a better transcultural understanding through.
MANISHA VERMA, VASUDEVA VARMA PATENT SEARCH USING IPC CLASSIFICATION VECTORS.
Jumping Off Points Ideas of possible tasks Examples of possible tasks Categories of possible tasks.
Evaluating the Performance of IR Sytems
1 Distance education : What could technology offer ? Gérard CHOLLET ENST/CNRS-LTCI 46 rue Barrault PARIS cedex 13
HYPERGEO 1 st technical verification ARISTOTLE UNIVERSITY OF THESSALONIKI Baseline Document Retrieval Component N. Bassiou, C. Kotropoulos, I. Pitas 20/07/2000,
Talk, Translate, and Voice By: Jill Gruttadauro, Amanda Swetish, Porter Waung.
 Official Site: facility.org/research/evaluation/clef-ip-10http:// facility.org/research/evaluation/clef-ip-10.
Advanced Auto attendant v3.0. December 2003 Page 2 New Auto Attendant Features for 3.0 Allow different languages on different dialogs New Language support.
CS344: Introduction to Artificial Intelligence Vishal Vachhani M.Tech, CSE Lecture 34-35: CLIR and Ranking in IR.
Why do we study English? Form 9, unit 6.
Richard Baraniuk International Experiences with Open Educational Resources.
CLEF Ǻrhus Robust – Word Sense Disambiguation exercise UBC: Eneko Agirre, Oier Lopez de Lacalle, Arantxa Otegi, German Rigau UVA & Irion: Piek Vossen.
PATENTSCOPE Patent Search Strategies and Techniques Andrew Czajkowski Head, Innovation and Technology Support Section Centurion September 11, 2014.
1 Formal Models for Expert Finding on DBLP Bibliography Data Presented by: Hongbo Deng Co-worked with: Irwin King and Michael R. Lyu Department of Computer.
“ SINAI at CLEF 2005 : The evolution of the CLEF2003 system.” Fernando Martínez-Santiago Miguel Ángel García-Cumbreras University of Jaén.
A Simple Unsupervised Query Categorizer for Web Search Engines Prashant Ullegaddi and Vasudeva Varma Search and Information Extraction Lab Language Technologies.
Concept-based Image Retrieval: The ARRS GoldMiner ® Image Search Engine Charles E. Kahn Jr., MD, MS Medical College of Wisconsin Milwaukee, Wisconsin,
Thanks to Bill Arms, Marti Hearst Documents. Last time Size of information –Continues to grow IR an old field, goes back to the ‘40s IR iterative process.
The PATENTSCOPE search system: CLIR February 2013 Sandrine Ammann Marketing & Communications Officer.
CROSSMARC Web Pages Collection: Crawling and Spidering Components Vangelis Karkaletsis Institute of Informatics & Telecommunications NCSR “Demokritos”
Applying the KISS Principle with Prior-Art Patent Search Walid Magdy Gareth Jones Dublin City University CLEF-IP, 22 Sep 2010.
The CLEF 2003 cross language image retrieval task Paul Clough and Mark Sanderson University of Sheffield
MIRACLE Multilingual Information RetrievAl for the CLEF campaign DAEDALUS – Data, Decisions and Language, S.A. Universidad Carlos III de.
Xiaoying Gao Computer Science Victoria University of Wellington Intelligent Agents COMP 423.
Péter Schönhofen – Ad Hoc Hungarian → English – CLEF Workshop 20 Sep 2007 Performing Cross-Language Retrieval with Wikipedia Participation report for Ad.
Nan Yang Chinese Terminologist Microsoft Language Excellence Shanghai, August 2008.
New RCLayout. Do product layout 3 improvements All products Local databases New functionalities.
5 th EI World Congress - Berlin, July 2007 Use of the Web and Internet Technologies to enhance Teacher Union Work.
Chapter 23: Probabilistic Language Models April 13, 2004.
From Text to Image: Generating Visual Query for Image Retrieval Wen-Cheng Lin, Yih-Chen Chang and Hsin-Hsi Chen Department of Computer Science and Information.
1 Flexible and Efficient Toolbox for Information Retrieval MIRACLE group José Miguel Goñi-Menoyo (UPM) José Carlos González-Cristóbal (UPM-Daedalus) Julio.
Medical retrieval 2008 New data set with almost 66,000 images Thirty topics were made available, ten in each of three categories: visual, mixed, and semantic.
Customization in the PATENTSCOPE search system Cyberworld November 2013 Sandrine Ammann Marketing & communications officer.
The Cross Language Image Retrieval Track: ImageCLEF Breakout session discussion.
MobiWeb2.0 Outreach Activities Dr. Marie-Claire Forgue Head of W3C European Communications MobiWeb2.0 Outreach Work Package Leader 1.
Combining Text and Image Queries at ImageCLEF2005: A Corpus-Based Relevance-Feedback Approach Yih-Cheng Chang Department of Computer Science and Information.
Cross Language Information Exploitation of Arabic Dr. Elizabeth D. Liddy Center for Natural Language Processing School of Information Studies Syracuse.
Xiaoying Gao Computer Science Victoria University of Wellington COMP307 NLP 4 Information Retrieval.
Analysis of Experiments on Hybridization of different approaches in mono and cross-language information retrieval DAEDALUS – Data, Decisions and Language,
Usefulness of Quality Click- through Data for Training Craig Macdonald, ladh Ounis Department of Computing Science University of Glasgow, Scotland, UK.
Factiva.com. What is Factiva? Joint venture between two of the world’s leading sources of company and business news + Knight Ridder Media General Hoover’s.
Find International Driving Document Translator Online
Free Powerpoint Templates Page 1 Free Powerpoint Templates Pass Microsoft Exam MCSE: Data Platform.
Using the Automatic Captions Feature. Objectives Learn how to use the Automatic Captions feature in YouTube  Edit the generated captions  Extract the.
PATENTSCOPE Patent Search Strategies and Techniques Andrew Czajkowski Head, Innovation and Technology Support Section.
1 SINAI at CLEF 2004: Using Machine Translation resources with mixed 2-step RSV merging algorithm Fernando Martínez Santiago Miguel Ángel García Cumbreras.
Pass Microsoft MCSE Exam MCSE: Business Intelligence
RECENT TRENDS IN SMT By M.Balamurugan, Phd Research Scholar,
Designing Cross-Language Information Retrieval System using various Techniques of Query Expansion and Indexing for Improved Performance  Hello everyone,
HiRISE: HiTranslate Project Marks One Year of Success
Oracle Supplier Management Solution Product Availability
CLIR PATENTSCOPE search system
Trust and Culture on the Web

PolyAnalyst Web Report Training
COUNTRIES NATIONALITIES LANGUAGES.
Claro ScanPen Reader By Claro Software Limited
Introduction to Search Engines

Active AI Projects at WIPO
Presentation transcript:

UA in ImageCLEF 2005 Maximiliano Saiz Noeda

Index System  Indexing  Retrieval Image category classification  Building  Use Experiments and results R2D2 Joint participation Conclusion & future work

System Based on textual captions  11(13) languages: Chinese (simplified and traditional), Dutch, English, French, German, Greek, Italian, Japanese, Portuguese, Russian and Spanish (europ. and latinoam.)  Web translator investigation for better performance on 2004 dataset Combination of probabilistic and automatic extracted knowledge Use of Xapian, a probabilistic IR engine.  Automatic feedback using first 23 relevant documents Two phases  Indexing (offline)  Retrieval (online)

System: Indexing Creation of three indexes  Stems  Words  Stem bigrams Token weighting depending on:  Weight of field where token is contained  Upper or lowercase in token first letter

System: Retrieval Three retrievals, one in each index and combine them In each retrieval:  First basic retrieval  Feedback  second enriched retrieval  Add category information  Combine three results and sort

System: Retrieval (II) Combination of three indexes to improve performance  Assigning weights to each retrieved list  Obtaining their mean average

Image category classification Knowledge base to enrich IR Automatic created from St. Andrews Corpus Extract categories related with a query and retrieve documents with the same categories

Image category building Create a document with words that appear with a concrete category

Image category use Queries are posed to category index to retrieve a list of “category-documents”

Experiments and Results Different features combined to create more than 100 experiments stemwordbigramcatsfeed 2005 MAPPrec. 10Recall Baseline X ,51430,8377 Experiment1 XXX ,55360,8399 Experiment2 XXXX ,54290,8246 Experiment3 XXXXX ,54290,8246 BEST ,41350,550,8434

R2D2 Joint participation UA, SINAI (Jaén) and UNED (Madrid)  UA+SINAI English, Dutch, French, German, Italian, Russian and Spanish  UA+SINAI+UNED Spanish Voting method based on combining document weights from three systems  weight normalization (document weight/max. weight)  weight adding (for each document in each list)  priority for documents in all the lists

Joint participation: results LanguageUASINAIUNEDUA-SINAIUA-SINAI-UNED English0.3966(14)0.3727(30) (7)- Dutch0.2765(8)0.3397(2) (1)- French0.2621(6)0.2864(1) (5)- German0.2854(7)0.3004(4) (1)- Italian0.2230(4)0.1805(11) (2)- Russian0.2683(3)0.2229(11) (5)- Spanish (eur)0.2105(12)0.2416(5)0.3175(1)0.2668(4)0.3020(2) Spanish (lat)0.3179(2)0.2967(8)0.2585(17)0.3447(1)0.3054(4)

Thank you very much!