NDAK20002U Neural Information Retrieval (NIR)

Volume 2020/2021
Education

MSc Programme in Computer Science

Content

 

The course objective is to offer an advanced introduction into information retrieval. The goal is to understand and model how people search for, access and use information, in order to design and evaluate reliable retrieval algorithms. Through realistic and sound projects, the course aims to stimulate and prepare students for their MSc thesis work.

 

The course will focus on these main questions:

  • How can we design efficient retrieval systems?
  • How can we design effective retrieval systems?

 

Content in detail:

Architecture of an IR system

  • Basic building blocks
  • Crawling, filtering and storing information
  • Ranking with indexes

 

Information ranking models

  • Deep learning for search engines
  • Probabilistic & machine learning models
  • Complex queries and combining evidence
  • Domain-specific ranking
  • Evaluation and optimisation

 

Learning Outcome

Knowledge of

  • The basic architecture of retrieval systems
  • The basic models and techniques for collecting, storing and ranking information
  • Different criteria for information retrieval evaluation

 

Skills in

Students should be able to transfer the above knowledge to real-world tasks by:

  • Designing appropriate strategies for crawling, storing and ranking information
  • Planning and carrying out appropriate evaluations

 

Given a working retrieval system, students should be able to:

  • Diagnose problems in its main information processing functions
  • Design and calibrate appropriate solutions

 

Competences to

  • Explain the basic information retrieval principles to both laymen and specialists
  • Use standard procedures and practices when designing or implementing information retrieval solutions
  • Present evaluation analyses and results in a proper format of a written report such that a technically qualified person can follow and obtain similar findings

 

Literature

The literature consists of seminal research and review articles from central journals and selected papers from peer-reviewed conferences, textbooks and research reports. This is supplemented with practical experience gained through lab sessions.

The literature will be listed in Absalon.

It is expected that students know how to program and have a working knowledge of Machine Learning corresponding to the course Machine Learning (ML) or an equivalent course.

Academic qualifications equivalent to a BSc degree is recommended.
The course will use a combination of lectures, lab sessions, class discussions and student presentations. Where possible, relevant guest lecturers will be involved.
The course is identical to NDAK13001U Information Retrieval and Interaction.
It is not allowed to pass both courses.
  • Category
  • Hours
  • Lectures
  • 28
  • Preparation
  • 14
  • Theory exercises
  • 57
  • Practical exercises
  • 57
  • Project work
  • 50
  • Total
  • 206
Oral
Feedback by final exam (In addition to the grade)
Peer feedback (Students give each other feedback)
Credit
7,5 ECTS
Type of assessment
Portfolio
The elements included in the exam are:
(i) submission of the student’s own project report, and
(ii) oral presentation of the student's own project.
Aid
All aids allowed
Marking scale
7-point grading scale
Censorship form
No external censorship
Several internal examiners
Re-exam

For the re-exam the student must submit a new projects/assignments no later than three weeks before the re-exam date.

Additionally an 20-minute oral examination without preparation will be administered covering the full course syllabus.  All aids are allowed for the oral examination.

The part-examinations/assignments must be individually approved. The final grade is based on an overall assesment.

Criteria for exam assesment

See Learning Outcome.