FACTS-IR Workshop @ SIGIR 2019
This workshop explores challenges in responsible information retrieval system development and deployment. The focus will be on determining actionable research agendas on five key dimensions of responsible information retrieval: fairness, accountability, confidentiality, transparency, and safety. Rather than just a mini-conference, this workshop will be an event during which participants will also be expected to work. The workshop aims to bring together a diverse set of researchers and practitioners interested in helping to develop the technical research agenda for responsible information retrieval.
The full-day workshop will be organized in two parts:
- The first half-day will include presentations, based on the accepted papers by our PC members
- For the second half-day, participants will be organized in working groups and tasked with articulating a research agenda for one of the FACTS-IR topics (Fairness, Accountability, Confidentiality, Transparency, and Safety)
To ensure that the working groups in the second part of the workshop will be productive, we may cap the number of participants.
You can find the non-archival proceeding for the FACTS-IR workshop at: https://arxiv.org/html/1907.05755
9:00-9:10 Welcome (10)
9:10-9:25 Invited talk - Preethi Lahoti (MPI-INF): Operationalizing Individual Fairness for Algorithmic Decision Making (15)
9:25-9:35 Paper talk - Fairness in Recommendation Ranking through Pairwise Comparisons (10)
9:35-9:50 Invited speaker - Maria Panteli (BBC) (15)
9:50-10:00 Paper talk - CobWeb: A Research Prototype for Exploring User Bias in Political Fact-Checking (10)
10:00-10:10 Paper talk - Do Transformer Attention Heads Provide Transparency in Abstractive Summarization? (10)
10:10-10:20 Paper talk - Explaining Predictions from Tree-based Boosting Ensembles (10)
10:20-10:30 Paper talk - Global Aggregations of Local Explanations for Black Box models (10)
11:00-11:15 Invited talk – Krisztian Balog (University of Stavanger & Google) (15)
11:15-11:30 Invited talk - Mahmoud F. Sayed (University of Maryland) (15)
11:30-11:40 The FACTS of Technology-Assisted Sensitivity Review (10)
11:40-11:55 Invited talk - Pierre-Nicolas Schwab (RTBF) (15)
Breakout groups and goal setting
11:55-12:10 Overview of the afternoon session (15)
12:10-12:30 Breakout group assignment (20)
1:30-3:00 Breakout groups
3:30-3:50 Breakout groups wrap-up
3:50-4:40 Group discussion
4:40-5:00 Next steps and farewell
Call for Papers
The purpose of the workshop is to identify gaps in the technical emerging work on responsible IR, including undertheorized and underspecified issues related to each of these five areas of focus. We aim to create actionable technical research agendas for each of them.
To this end, we welcome technical contributions and position papers as either long papers (8 pages) or extended abstracts (2-4 pages) on a wide range of topics, including:
The IR system should avoid discrimination across people and communities. To do so the notion of fairness should be contextual and well grounded in the application setup and domain. Achieving fairness may be further complicated by the multi-stakeholder nature of most IR systems.
The IR system should be able to justify its recommendations or actions to users and other stakeholders, as well as be reliable at all times. This requires an understanding of the potential harms of using the system and of who is more likely to be affected. It also requires recourse avenues and processes for redress.
The output or actions of the IR system should not reveal secrets. For example, IR systems often combine extensive behavioral logs to model their users, which if not properly handled can result in unintended leakage of information.
The IR system should be able to explain why and how the suggested results were obtained to users and other interested stakeholders. Providing proper explanations may require to answer who the users and the stakeholders are. More broadly, the IR systems should be able to enable third parties to monitor and probe that they behave as expected.
The IR system should be resilient to manipulation by possible adversarial parties, and should not expose the users to undesirable, harmful content.
FACTS-IR is interested in technical contributions on these topics.
- May 12, 2019: Workshop paper submissions due
- May 31, 2019: Workshop paper notifications
- June 30, 2019: Upload final version to arXiv
- July 25, 2019: Workshop Day at SIGIR 2019
How to submit
Papers should be in English, in PDF, and formatted using the standard ACM sigconf format, using the template available at the ACM site; Word users should use the Interim Template. FACTS-IR uses the HotCRP submission system: FACTS-IR submission site.
No official proceedings will be published, however accepted long paper submissions will be expected to be posted on arXiv.org. This is optional for extended abstract submissions. The review process is a single-blind.
- Alexandra Olteanu (Microsoft Research, US & Canada)
- Jean Garcia-Gathright (Spotify, US)
- Maarten de Rijke (University of Amsterdam, The Netherlands)
- Michael Ekstrand (Boise State University, US)
- Ana-Andreea Stoica (Columbia University, US)
- Asia Biega (MPI, Germany)
- Ashudeep Singh (Cornell University, US)
- Avishek Anand (L3S, Germany)
- Carlos Castillo (UPF, Spain)
- Christo Wilson (Northeastern University, US)
- Damiano Spina (RMIT, Australia)
- Daniel Kluver (University of Minnesota, US)
- Diane Kelly (University of Tennessee, US)
- Dong Nguyen (Alan Turing Institute, UK)
- Emilia Gómez (UPF, Spain)
- Emre Kiciman (Microsoft Research, US)
- Faegheh Hasibi (Radboud University Nijmegen, The Netherlands)
- Fernando Diaz (Microsoft Research, Canada)
- Gianluca Demartini (The University of Queensland, Australia)
- Hinda Haned (Ahold Delhaize & University of Amsterdam, The Netherlands)
- Ingmar Weber (QCRI, Qatar)
- James Thom (RMIT, Australia)
- Mark D. Smucker (University of Waterloo, Canada)
- Meike Zehlike (TU Berlin, Germany)
- Min Zhang (Tsinghua University, China)
- Pierre-Nicolas Schwab (Solvay Brussels School of Economics and Management, Belgium)
- Rishabh Mehrotra (Spotify, UK)
- Ronald Robertson (Northeastern University, US)
- Solon Barocas (Cornell University & Microsoft Research, US)
- Stefano Balietti (Microsoft Research, US)
- Suzan Verberne (Leiden University, The Netherlands)
- Toshihiro Kamishima (National Institute of Advanced Industrial Science and Technology, Japan)
Please mail us at firstname.lastname@example.org if you have any questions about the workshop.