About

The 21st Dutch-Belgian Information Retrieval (DIR) Workshop is a forum where researchers and practitioners in Information Retrieval come together to connect, exchange ideas, and discuss the latest developments in IR and related research areas. DIR has an established tradition dating back to the year 2000, making DIR2023 the 21st edition. Please find an incomplete list of earlier editions here

Location

Next Delft Molengraaffsingel 8 2629 JD Delft

Registrations now open at:

Present Your Work

At this full-day event there will be plenty of opportunities for students, academics, and industry professionals to interact and learn from one another. In DIR 2023, we are excited to continue the dialogue on diversity and inclusion within the IR community. DIR 2023 will also feature special sessions shining a spotlight on the region's ongoing research. We invite you to submit contributions in three categories

1. Published work: Share your research papers published in the past year at IR-related venues like SIGIR, ECIR, ACM TWeb, and more.
2. Research spotlight: This category aims to spotlight emerging research directions from young academics, i.e., assistant professors, in the region.
3. Resources: This refers to demos, datasets, experimental platforms (already presented at other venues) that are of interest to the IR community. We also welcome European projects seeking exposure among workshop attendees, ending this event, as well as aiming to foster new partnerships, and for recruiting purposes.

Please submit your contributions by submitting the form here

Please note that the special sessions have limited presentation slots available. Our goal is to curate a diverse program that reflects the breadth of IR research across the Belgian and Dutch communities. Accepted contributions in the published work and research spotlight categories will receive a 15-minute oral presentation (12 minutes for presentation and 3 minutes for Q&A), while resources will be presented as posters.

Important dates

Nominate your contribution by: October 23, 2023
Notification of contributions to spotlight during DIR 2023: October 30, 2023
DIR 2023: November 27, 2023

Information Retrieval in the age of LLMs

''In this talk I will discuss the relationship between Information Retrieval (IR) and Large Language Models (LLMs). Generative LLMs can help IR with data augmentation, query rewriting, summarisation, or even directly be used to generate document ids in ranking. Vice versa, IR helps LLMs with source retrieval in retrieval-augmented generation. But IR has more to offer. IR is a user-oriented field: we want to help users solve a problem. We think critically about evaluation metrics and multiple dimensions of evaluation: we do not only care about relevance but also about diversity, bias, and explainability. In chatbots, IR can help the reliability and trustworthiness of the returned information. I will talk about these opportunities and call for a central role of IR in the realm of LLMs."

Short Bio

Suzan Verberne is professor of Natural Language Processing (NLP) in the Leiden Institute of Advanced Computer Science at Leiden University. She obtained her PhD in 2010 on the topic of Question Answering and has since then been working on the edge between NLP and Information Retrieval (IR). She has supervised projects involving a large number of application domains: from social media to law and from archaeology to health. Her research focus is to advance NLP "beyond the benchmark", addressing challenging problems in specific domains. This involves the development of Large Language Models for low-resource settings, and addressing issues related to misinformation, transparency, and bias. Suzan is active in the NLP and IR communities. Notably, she is co-chair of the SIGIR Virtual Forum initiative, and Diversity-Equity-Inclusivity (DEI) co-chair for the worldwide SIGIR organisation.https://liacs.leidenuniv.nl/~verbernes/

On relevance

(Slides at : Here)

The concept of 'relevance' is essential to much work in IR. It grounds what will be successful to human users, and what thus needs to be prioritized.However, when striving for diversity, one may need to look beyond that what traditionally has been deemed relevant. As for this, in my talk, I will give several examples of ways in which I pursued routes in my career that at first sight would be deemed irrelevant or even irrational, but that to me were essential in ensuring that broader perspectives are identified, heard, and kept within reach."

Short Bio

Dr Cynthia C. S. Liem MMus is an Associate Professor in the Multimedia Computing Group of Delft University of Technology, and pianist of the Magma Duo. Her research interests are in trustworthy and responsible AI; here, she especially focuses on techniques that make people discover new interests and content which would not trivially be retrieved, and questions of validation and validity in data-driven decision-making. After starting in music information retrieval, today, her research considers broader public-interest domains with high societal impact. She initiated and co-coordinated multiple European research projects on technological enrichment of digital musical heritage, and gained industrial experience at Bell Labs Netherlands, Philips Research and Google. Cynthia received recognition through multiple awards (Lucent Global Science and Google Anita Borg Europe Memorial scholarships, Google European Doctoral Fellowship, finalist of the New Scientist Science Talent Award, Researcher-in-Residence at the National Library of The Netherlands), is a member of the Dutch national Young Academy, and still performs as a musician. With her track record in inter- and transdisciplinary collaborative research and public engagement, she became a sought-after educator and communicator, and is a core team member of TU Delft | The Hague, seeking to better connect engineering expertise to policymaking.