Open positions

PhD

No open positions

Master’s theses

Analyzing Intersectional Bias in LLM-based Recommender Systems.

Description: Recently, Recommendation Systems (RSs) have leveraged Large Language Models (LLMs) to enhance personalized recommendations. However, LLMs can perpetuate social biases, raising concerns about their trustworthiness in RSs. RSs are widely used in sectors like job markets, finance, and medicine that critically need fair decision-making. While fairness in traditional RSs has been explored, the trustworthiness of LLM-based RSs remains understudied. Most existing research focuses on bias related to specific group identities such as race or gender. However, recent work emphasizes the need to address “Intersectional Fairness”, where interactions across multiple identity dimensions result in unique discrimination for subgroups. This project aims to comprehensively analyze intersectional biases in both traditional and LLM-based RSs, with a focus on critical applications such as news recommendations.
Requirements: Mining of Social Data, Deep Learning, Advanced Probabilistic Machine Learning (meriting)
Contact: Ece Calikus

Exploring the Power of LLM-Generated Fake News with the Focus on Scientific Discussions

Description: Advances in Large Language Models (LLMs) have enabled malicious actors to generate fake content that mimics the style of trustworthy news sources. Recent studies have revealed that LLM-generated fake news can closely resemble reliable sources of information, reducing the effectiveness of state-of-the-art text-based misinformation detection tools. The rise of misinformation and pseudoscience in scientific discussions on topics such as vaccines, medicine, and climate change is an important societal problem. These beliefs, theories, or practices are disguised as scientific but lack empirical evidence, rigorous methodology, and adherence to the scientific method. They can mislead individuals by presenting unverified or disproven claims as factual, undermining legitimate scientific practices. This project will implement a prompt-based approach to generate realistic fake content related to scientific topics and investigate whether state-of-the-art methods can detect them.
Requirements: Mining of Social Data, Deep Learning, Advanced Probabilistic Machine Learning (meriting)
Contact: Ece Calikus

Interpretable Anomaly Detection for Tabular Data Using Diffusion Models

Description: This project aims to develop interpretable anomaly detection techniques for tabular data using diffusion models, leveraging their robust generative capabilities to accurately model complex data distributions. Unlike traditional methods, diffusion models offer a unique approach by gradually transforming noise into structured data, enabling a precise representation of normal behavior. By learning this distribution, we will detect anomalies as significant deviations from the expected patterns. The focus on interpretability ensures that detected anomalies can be understood and contextualized, providing meaningful insights and explanations critical for decision-making in high-stakes domains such as finance, healthcare, and industrial monitoring.
Requirements: Mining of Social Data, Deep Learning, Advanced Probabilistic Machine Learning (meriting), Advanced Deep Learning for Image Processing (meriting)
Contact: Ece Calikus

Analyzing the Downstream Implications of Pretrained Image Representations in Visual Topic Models

Description: This project investigates how the choice of pretrained image representation influences the performance and interpretability of visual topic models, focusing on feature extractors trained with different objectives (e.g., language-supervised, self-supervised, or vLLM’s). It examines the impact on model fit, topic coherence, and the labels assigned to discovered topics.
Requirements: Mining of Social Data, Introduction to Image Analysis (meriting), Advanced Deep Learning for Image Processing (meriting), Advanced Probabilistic Machine Learning (meriting)
Contact: Matias Piqueras

A comparative study of fairness-aware community detection methods

Description: With the growing interest in the field of algorithmic fairness, various approaches have been recently introduced to partition graphs into clusters while simultaneously satisfying fairness constraints. This project aims to study and experimentally compare these methods, their scalability, as well as the effect of different demographic fairness functions.
Requirements: Participation in course: Mining of Social Data
Contact: Georgios Panayiotou

Bachelor theses

No open positions

Before joining

This section collects some information about life in Sweden and about how we work, that can be useful if you are interested in joining the lab as a PhD student or postdoc.

Working and living in Sweden

Sweden is a fantastic place for living and working. Swedes are friendly and speak excellent English. The quality of life is high, with a strong emphasis on outdoor activities. The Swedish working climate emphasises an open atmosphere, with active discussions involving both junior and senior staff. Spouses of employees are entitled to work permits. Healthcare is free after a small co-pay and the university subsidises athletic costs, such as a gym membership. The parental benefits in Sweden are among the best in the world, including extensive parental leave (for both parents), paid time off to care for sick children, and affordable daycare. For more information, be sure to read Why choose Sweden? and Why choose Uppsala University?.

Expected availability from Infolab members

Given the creative nature of part of our work, Infolab members are generally free to work when and where it best fits them, compatibly with tasks requiring presence (e.g., teaching) and regulations that may be enforced by the Department or University.

At the same time, everyone is expected to contribute to the lab, which implies being available to give feedback, sharing their current research with other members (see below), and participating in lab meetings.

Lab members are free to take time to work offline. In particular, it is not necessary to be continuously online on our Slack space. Checking the Slack at least once a day is however expected, and if one cannot be online for a few days for any reason (e.g. holidays) it is good practice to inform the other lab members.

Research quality at the Infolab

While “quality” is a subjective term, and we expect lab members to reflect about this concept and adopt their own definitions of quality, we also have some guidelines to ensure that some common ground is established. In particular, we want our research to be based on appropriate methods, clearly and transparently communicated, developed with ethical and societal awareness, replicable (when possible), and usable by others.

To achieve this, research at Infolab is organised into research activities. A research activity is anything leading to a published article, a grant proposal, a PhD dissertation, an undergraduate thesis, research software, etc. Each research activity is led by a member of the lab, who is the contact author and is responsible for the research to advance and for its quality. Under normal circumstances, the contact author should be available to answer questions about the activity and its outcomes for at least two years after the outcome has been made public, even in case of a change of affiliation.

Members of the lab should feel a responsibility of getting and receiving feedback from other members. At the beginning of a new research activity, the (contact) author(s) is encouraged to pitch the idea to other lab members, who should be available to provide constructive feedback. Before submitting a paper or research proposal the authors are invited to share it with the other lab members, who should be available to provide constructive feedback. Getting feedback even earlier is also encouraged, although not expected.

At submission time, for all products containing experiments the code to replicate the experiments should be made available in a repository in the Infolab’s git workspace, in an easy-to-execute format (with exceptions, for example double-blind review processes).

Authorship

Each research activity must have a clear list of authors. At the Infolab we define authors as in: “each author is expected to have made substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data; or the creation of new software used in the work; or have drafted the work or substantively revised it; AND to have approved the submitted version (and any substantially modified version that involves the author’s contribution to the study); AND to have agreed both to be personally accountable for the author’s own contributions and to ensure that questions related to the accuracy or integrity of any part of the work, even ones in which the author was not personally involved, are appropriately investigated, resolved, and the resolution documented in the literature.”1

Start-up checklist

When joining the lab:

  1. If you are not already in Uppsala, make sure you start looking for an accommodation well in advance, this can take time.
  2. A current faculty member will register you to our internal systems.
    • Slack (we do not normally use email for communication).
    • The lab’s Git space (please provide your git account).
    • Group storage space.
  3. Follow the LinkedIn, X, and Facebook accounts of the lab, if you use these social media for work.

  1. Transparency in authors’ contributions and responsibilities to promote integrity in scientific publication. Marcia K. McNutt, Monica Bradford, Jeffrey M. Drazen, Brooks Hanson, BobHoward, Kathleen Hall Jamieson, Véronique Kiermer, Emilie Marcus, Barbara Kline Pope, Randy Schekman, Sowmya Swaminathan, Peter J. Stang, Inder M. Verma. Proceedings of the National Academy of Sciences Mar 2018, 115 (11) 2557-2560; DOI:10.1073/pnas.1715374115 ↩