Course in Advanced Information Retrieval

General Information

Lecturer Jun.-Prof. Dr. Martin Potthast
Lab Advisors Theresa Elstner
Workload 2 SWS Lecture, 3 SWS Lab
Lecture Tuesdays, 11:15 - 12:45 - Paulinum P801 video lectures
Lab Tuesdays, 13:15 - 14:45 - Seminargebäude 3-10 BBB
Contact Email, or via Discord server "irlecture"
Exam To be announced.

Announcements

  • Final Presentations on February, 1st 2022 9:00 - 11:00 h.
  • Changed required template format for final report: [Latex Template]

Organization

  • Some Lectures are additionally available as videos. The videos can be accessed by following the lecturenotes below, or on the Webis youtube channel. [playlist]
  • Lab participation is a prerequisite to complete the module.
  • Communication
    • Lecture website - materials and announcements will be uploaded on this website.
    • Discord - there is a dedicated Discord server for this lecture. Check your mails for an access code. There are different channels for questions regarding lecture and lab, group finding, and each lab group will get a dedicated channel for internal communication. Please join the server and choose a Nickname such that we can identify you (at least surname).
    • Email - important announcements will be sent out via mail.

Lecturenotes

Lab Project

The lab consists of building and evaluating an information system for a specific domain. This entails related work search, data cleansing, indexing, selection and implementation of suitable retrieval models, evaluation of search quality, and the submission of a written report and well-documented source code.

Lab Lecturenotes

Lab Classes

  • 2021-10-27. Introduction [slides]
  • 2022-01-25. Final Presentations
    Time Group
    11:00 Korg
    11:30 Pearl
    12:00 Goldar
    12:30 Boromir
  • 2022-02-01. Final Presentations
    Time Group
    09:00 Hit-Girl
    09:30 Porthos
    10:00 Jester
    10:30 Aramis

Lab Material

  • Literature
    • Overview
      • Bondarenko et al. Overview of Touché 2021: Argument Retrieval. (CLEF 2021). [link]
    • Task 1
      • Wachsmuth et al. Building an Argument Search Engine for the Web (ArgMining 2017). [link]
      • Ajjour et al. Data Acquisition for Argument Search: The args.me corpus. (KI 2019). [link]
      • Potthast et al. Argument Search: Assessing Argument Relevance. (SIGIR 2019). [link]
      • Wachsmuth et al. Computational Argumentation Quality Assessment in Natural Language. (EACL 2017). [link]
    • Task 3
      • Kiesel et al. Image Retrieval for Arguments Using Stance-Aware Query Expansion. (ArgMining 2021). [link]
      • Dimitrov et al. SemEval-2021 Task 6: Detection of Persuasion Techniques in Texts and Images. (SemEval 2021). [link]
      • Yanai Image collector III: a web image-gathering system with bag-of-keypoints. (WWW 2007). [link]
  • Data
    • Data is organized in a dedicated Git repository [link]
  • Example (Argument Search)
    • Demo: [args.me]
    • Source: [Git]
    • Note that you do not need to implement a frontend in the lab. Your system only has to interact with Tira.
  • TIRA
    • Check your groups Discord channel for credentials!
    • Quickstart Guide [quickstart]
    • Video Tutorial [video]
    • Tutorial Code [git]

Lab Report

A written report is expected at the end of the semester. Note that grades will be based on the lab report and on the final presentation. Detailed information on what is expected of the report and the talk can be found below.

  • Language: English or German
  • Structure:
    1. Introduction
    2. Related Work
    3. Methodological Approach (i.e., a description of your contribution)
    4. Evaluation
    5. Discussion and Conclusion
  • Style: please use the following template for your report [Latex Template]
  • Length: minimum 10 pages.
  • Supplementary Material:
    • your source code in a git repository with full commit history
    • a working deployment of your system on the Tira platform
  • Due Date: reports have to be turned in on 28.02.2022
  • Talk: in addition to the report, we will have short (~20min + questions) talks with each group at the end of the semester. Each person of the group should have their fair share of talking and the talk should cover all important aspects of your approach.