Information Retrieval - Principles, Methods, and Systems

Exam Grades

The exam grades (Matrikelnummer and Grade) can be accessed here using your n.ethz login and password.

Content

The course presents an introduction to the field of information retrieval and discusses automated techniques to effectively handle and manage unstructured and semi-structured information. This includes methods and principles that are at the heart of various systems for information access, such as Web or enterprise search engines, categorization and recommender systems, as well as information extraction and knowledge management tools.

Exam

Time: Wednesday, Dec 12th, 09:00am. Please be there at 09:00am.
Place: IFW A36.
Duration: 1 hour.
Type: Open book (you can use any materials).
Language: English.
Please bring your Student ID with you.

A template exam is available here (use your n.ethz user and password).

Lecture Notes

• Lecture 1: Introduction
• Lecture 2: Indexing (1) [updated 11.10.2007]
• Lecture 3: Indexing (2) [updated 11.10.2007]
• Lecture 4: Retrieval Models (1) [updated 23.10.2007]
• Lecture 5,6: Retrieval Models (2)
• Lecture 7: Hypertext & Web Retrieval
• Lecture 8: Text Categorization
• Lecture 9: Feedback & Interaction
• Lecture 10: Recommender Systems & Social Filtering

Exercise Sheets

Syllabus:

1. Introduction (90' – Lecture 1)

• organizing digital information
• brief history of information retrieval
• retrieval systems & applications
• basic retrieval architecture
• challenges in information retrieval & overview

2. Document Indexing (120' – Lecture 2 & 3)

• metadata and controlled vocabulary
• automatic full-text indexing, tokenization & term processing
• statistical properties of term occurrences
• data structures for dictionaries
• inverted files & index compression
• index construction
• in-memory serving systems

3. Retrieval Models (150' – Lecture 3 & 4)

• Boolean retrieval model
• vector space model
• latent semantic indexing
• probabilistic retrieval models
• language model paradigm in information retrieval
• probabilistic latent semantic indexing
• relevance-based models
• query processing
• evaluation of retrieval systems

4. Web Retrieval (180' – Lecture 5 & 6)

• properties of the web as a graph
• query-free quality measures: PageRank
• co-citation analysis & bibliometrics
• hyperlink induced topic search (Hubs & Authorities)
• trust and distrust
• spam detection
• spell checking & query expansion

5. Recommender Systems (180' – Lecture 7 & 8)

• problem setting & applications, Netflix
• memory-based methods: GroupLens
• matrix-based methods
• probabilistic methods
• advanced topics

6. Automatic Text Categorization (150' – Lecture 9 & 10)

• taxonomies and their applications
• naive Bayes classification
• linear classification (1): perceptron and variants
• linear classification (2): support vector machines
• practical aspects

7. Feedback & Usage Data (120' – Lecture 10 & 11)

• relevance feedback
• pseudo relevance feedback
• click-based feedback
• learning to rank

8. Information Extraction (90' – Lecture 12)

• named entity recognition
• hidden Markov models
• wrapper induction
• extraction pattern learning

9. Multimedia Retrieval (90' – Lecture 13)

• similarity-based retrieval
• nearest neighbor queries
• content-based image & video retrieval systems
• relevance feedback
• retrieval models for text and images

10. Interactive Retrieval (90' – Supplementary Lecture)

• TBD

Total = 13-14 lectures

Lecturer

Thomas Hofmann (Google)

Class Schedule

2V1U

Language

English

URL

http://www.dbis.ethz.ch/education/ws0708/inf_retrieval