Information Retrieval

Credits
6
Types
Elective
Requirements
This subject has not requirements , but it has got previous capacities
Department
CS
The amount of information stored digitally in organizations, or collectively on the web, is today large enough to make searching this information a generally complicated task. The field known as "Information Retrieval" finds methods to organize information in such a way that finding information afterwards can be done simply and efficiently. We will cover basic keyword-based techniques to search in textual information. Then, we will examine search in the web, where hyperlinks can be used not only to direct the search but to assess the interest value of each page - as is the case with the well-known PageRank algorithm. We will see extensions of these techniques to the case of Social Networks where interactions among users can provide very useful information. Finally, we will study how the tecnologies known as Big Data and recommendation complement Information Retrieval techniques in contemporary systems.

Teachers

Person in charge

Others

Weekly hours

Theory
2
Problems
1
Laboratory
1
Guided learning
0
Autonomous learning
4

Contents

  1. Introduction
    Need of search and analysis techniques of massive information. Search and analysis vs. databases. Information retrieval process. Preprocessing and lexical analysis.
  2. Models of information retrieval
    Formal definition and basic concepts: abstract models of documents and query languages. Boolean model. Vector model. Latent Semantic Indexing.
  3. Implementation: Indexing and searching
    Inverse and signature files. Index compression. Example: Efficient implementation of the rule of the cosine measure with tf-idf. Example: Lucene.
  4. Evaluation in information retrieval
    Recall and precision. Other performance measures. Reference collections. Relevance feedback and query expansion.
  5. Web search
    Ranking and relevance in the web. The PageRank algorithm. Crawling. Architecture of a simple web search system.
  6. Architecture of massive information processing systems
    Scalability, high performance, and fault tolerance: the case of massive web searchers. Distributed architectures. Example: Hadoop.
  7. Network analysis
    Descriptive parameters and characteristics of networks: degree, diameter, small-world networks, among others. Algorithms on networks: clustering, community detection and detection of influential nodes, reputation, among others.
  8. Information Systems based on massive information analysis. Combination with other technologies.
    Search Engine Optimization. Joint use of IR techniques with Data Mining and Machine Learning. Recommender Systems.

Activities

Activity Evaluation act


Theoretical development of topics 1 to 8 of the course

The student will attend the instructor's presentation and actively participate in the initial discussion of the challenge to be solved in that session.

Contents:
Theory
26h
Problems
0h
Laboratory
0h
Guided learning
0h
Autonomous learning
26h

Exercises on topics 1 to 8 of the course

In each session, the instructor proposes a number of exercises (say, 4 to 7) on the topic just covered in theory. Next, a few of the problems (say, 3) are solved jointly. Students must solve the rest of the exercises and deliver them by the start of next session. A part of the session is devoted to discussing the possible questions that may have appeared while solving the problems pending from the last session.

Contents:
Theory
0h
Problems
13h
Laboratory
0h
Guided learning
0h
Autonomous learning
26h

Laboratory work on topics 1 to 8

The teacher will describe a practical work to be carried out related with the topics most recently covered. This may be a data analysis task, the implementation of an algorithm seen in class, or proposing a solution for an Information Retrieval scenario. The student completes the work as much as possible in class, although occasionally some additional time may be necessary. In many cases the student will have to produce a report on the work done and results obtained, to be delivered within some clearly stated deadline (say, 2 weeks).

Contents:
Theory
0h
Problems
0h
Laboratory
13h
Guided learning
0h
Autonomous learning
13h

Final exam

Final exam on the contents of the whole course

Week: 18
Theory
0h
Problems
0h
Laboratory
0h
Guided learning
0h
Autonomous learning
0h

Study and presentation of a scientific paper

Study and presentation of a scientific paper related to the course topic

Theory
0h
Problems
0h
Laboratory
0h
Guided learning
3h
Autonomous learning
10h

Teaching methodology

Sessions of theory + problemes of 3 sessions per week. The 2 hours of each session are theoretical expositions, and the third one is devoted to joint exercise solving. For each session, the student will have to deliver solutions to a few problems proposed but not solved in the previous session.

Laboratory sessions of 1 hour per week. For many of the sessons, the student will have to deliver a report of the work done and obtained results after about two weeks.

The working of each type of session is described in the "Activities" session.

Furthermore, at the end of the course each student must present to instructors and fellow students a scientific paper related to the course topic, in the format of a conference presentation. Near week 8 of the course, a list of papers will be made public, from which each student can choose one, or alternatively propose a paper of his/her choice, to be approved by the instructors. The date and time range for the presentations will be announced with at least 2 months time, and the schedule within the chosen day at least 1 week time.

Evaluation methodology

Define:

- NF as the grade of the final exam
- NE the grade of exercise assignments
- NL the grade of lab reports
- NA the grade from the presentation of a scientific article

(all in the range 0..10).

Then the final course grade is 0.3*NF + 0.25*NL + 0.25*NE + 0.2*NA.

Bibliography

Basic

Web links

Previous capacities

Those assumed for admision to MIRI plus those provided by the common learning phase.