faceted search

Design and Development of an Explainable AI (XAI) System for Item Ranking

Project Background

Explainable Artificial Intelligence (XAI) is a relatively new research field whose goal is to provide explanations of black-box models to help the users (developers and end-user) in understanding the behavior and decision-making mechanism of the models via tools, techniques, and algorithms.

Most of the current research in the XAI field has focused on the algorithmic part which has helped the developer in understanding how each layer in a neural network work or in mitigating algorithmic bias in people recommendation scenario. Few studies have taken into consideration how to visual communicate the explanation to the end-user and investigate the ranking problem.

Every time we use services like Galaxus, Zalando, or Amazon to buy an item we are prompt with a ranking of items but no explanation is given to us about why we are seeing those items and how well they aligned with our personal needs.

Project Goal

The ultimate goal of this Masters's Project is to design and develop an interactive human-centered explainable tool for item ranking. The main idea is to define a level of explanation (single item, group of items, entire ranking) and provide the user with the most suitable visual explanation that can support him\her in the decision process.

 

Work Tasks:

  1. Literature review of the existing human-centered explainable approach
  2. Identify requirements and guidelines to design an explainable tool to define a concise framework
  3. Design and develop a web-based item ranking explainer
  4. Conduct a user study using real-world datasets 

Organizational Information

The project will be supervised by Ibrahim Al Hazwani and Prof. Jürgen Bernard.

Ideally, the project should start at the end of April 2022 and the duration of the project will be no longer than 8 months. A group of ~3 students will be working on the project.

The project language is English, therefore, you are required to have good English knowledge (written and oral).

The programming languages of the project are not yet defined and are up for discussion by the project team (preferably Python or Java for the back end and JavaScript for the front end). The code will be versioned through GitLab and the tool will be hosted on the IFI server.

If possible (depending on the COVID situation), regular meetings will be held on-site.

The ultimate goal of the project would be to submit a paper at the end of 2022, therefore students with a research interest are preferred.

Your profile

You should have the following skills for this project:

  • Profound programming skills (preferably with Python, Java and/or JavaScript), ideally with full-stack background
  • Basic knowledge of databases (e.g. how to write a database statement to retrieve data)
  • Experience in working with version control systems (e.g. Git, we will be using GitLab)
  • Interest in research

In addition, it would be nice if you have experience in one of the following areas (but it is not a must):

  • UI / UX design
  • Testing (e.g. Unit Testing)
  • DevOps such as CI/CD and application deployment
  • REST or any other framework that supports the communication between the back and front end
  • CSS and HTML
  • Front end frameworks such as ReactJS, Angular, Material Design, Bootstrap
  • Back end frameworks such as Django, Flask, Spring

How to Apply / Contact

The topic will be assigned through the "Master Project Market" in Spring 2022. Watch out for the event sop that you don't miss it!

Interested students should send their complete application to Prof. Jürgen Bernard.

Applications should include:

  • Short, informal letter of motivation including your expectation for the project, the skills you want to learn, and your experience with data visualizations (if any)
  • CV
  • Transcript of records from your Master's and Bachelor's Degree

If you have any questions about the project, please feel free to contact  Prof. Jürgen Bernard or Ibrahim Al Hazwani.