DATA AND KNOWLEDGE

ENGINEERING GROUP

General Course Information

On this web page, information for the seminar 'Hot Topics in Information & Patent Retrieval', which is held during the Winter Term 2011/2012 by Farag Ahmed and Andreas Nürnberger, is provided.

Alle Informationen auf dieser Webseite sind auf Englisch angegeben, um auch englischsprachigen Studierenden die Teilnahme am Seminar zu ermöglichen. Vorträge und Ausarbeitungen können sowohl in Englisch als auch in Deutsch gehalten bzw. erstellt werden.

Description of topics

This seminar is devoted to emerging trends in Information Retrieval like Cross-language IR, Patent IR, Personalized IR and Opinion mining.

Cross-language IR (CLIR)

The increasing diversity of the Internet web sites has created millions of multilingual resources in the world wide web. Cross-language IR deals with retrieving information written in a language different from the language of the user's query. For example, users may write their queries in German but retrieve relevant documents written in French. Following topics are available:

• Cross language Information Retrieval: an overview of existing methods.
• CLIR interaction tools: analysis of appropriate user interfaces and interaction sequences.
• Multilingual Information Retrieval: an overview of issues and existing solutions to tackle them (e.g. for chinese , Arabic langauges etc.).
• Cross language Information Retrieval evaluations.

Patent information retrieval (patent IR)

Patent information retrieval (patent IR), sometimes called patent retrieval or patent search, is a sub branch of information retrieval that aims to support patent experts to retrieve patents that satisfy their information needs and search criteria. A common scenario in patent search is prior-art search, which is performed by patent experts to determine whether a new invention can be patentable. To perform their search tasks, patent experts use information retrieval systems and tools. Prior-art search can be achieved by considering all relevant information found in the patent data that can invalidate the novelty of a patent application claim. Thus, an invention is patentable only when no matched records for this patent claim can be found in the patent data. Following topics are available:

• Monolingual prior-art search approaches.
• Cross-lingual prior-art search approaches.
• Patent corpora available and their structure and properties.
• Short comparison about available patent retrieval systems.

Personalized IR

Current information retrieval systems return the same retrieval results to all users for identical queries, regardless of varied user interests and information needs. Personalized IR provides information for each user based on his/her interests .i.e., for a specific given query, personalized IR can provide different retrieval results for different users. Furthermore, personalized IR organizes retrieval results differently based on the user's interests, preferences, and information needs. Following topics are available:

• Sources of Personalization: Search History, Current Context, Collaborative approaches etc.
• Personalization of information retrieval through user profiling. User Modeling in Personalized Systems.
• Content and Collaborative-Based Personalization. Collaborative Filtering.
• Contextual Search.

 Opinion mining

An important part of our information-seeking behavior is to find out what other people think about specific topic. Opinion mining (OM) is a recent discipline at the crossroads of information retrieval and computational linguistics which is concerned not with the topic a document is about, but with the opinion it expresses. Opinion mining is used to identify and extract subjective information in source materials. Due to the abundance of opinion-rich resources that reside in online review sites, personal blogs etc. new challenges arise in how to use information technologies in order to search and understand the opinions of others. Following topics are available:

• Opinion mining and sentiment analysis. Construction of an Opinion/Review Search Engine.
• Linguistic resources for Opinion mining.
• Opinion summarization.

Semantic web

The core idea of the Semantic Web is to extend the current human-readable web by involving semantics to make the web easy understandable by machines. This includes the way of making the web data structured so that it can be interlinked and become more beneficial. This makes information easy to be connected and queried from different resource. Following topics are available.

• OWL: web ontology language.
• Web services and semantic web.
• Linked open data.

Course Schedule and Room Assignments

Course Staff

If you have any questions concerning the lectures or assignments please contact (if possible by email)

• Farag Ahmed
  E-Mail: farag.ahmed@ovgu.de
• Andreas Nürnberger
  E-Mail: andreas.nuernberger@ovgu.de

Requirements for the 'Schein'

The following criteria has to be fullfilled for the 'Schein':

• 20 minutes presentation (+10 minutes discussion) about an elected topic (topics can be elected in the first lesson): delivery of the slides 1 week before the presentation date.
• Active participation and reasonable discussion about the presentations.
• Write a scientific paper (hints will be given at the beginning of the seminar), 7-10 pages: delivery of the paper 1 week before the presentation date.
• participation in the other presentations.

If you need a 'benoteten Schein' you have to announce it in the first lesson!

Hints on the delivery

The delivery of the paper, the slides and the reports can be done by email to farag.ahmed@ovgu.de. Please send also the source file (word or latex document). The paper and the slides have to be send one week before the presentation date by wednesday at 8:00 am. The reports have to be send one day before the presentation date of the examined talk at 8:00 am. These deadlines must be strictly adhered.

Material

• Introduction (English)
• Springer LNCS-Style
• Example Paper