Integrity24: Workshop on Integrity in Social Networks and Media Mérida, Yucatán, MX Mérida, Yucatán, Mexico, March 8, 2024 |
Conference website | http://integrity-workshop.org/ |
Submission link | https://easychair.org/conferences/?conf=integrity24 |
Abstract registration deadline | January 22, 2024 |
Submission deadline | January 22, 2024 |
Workshop Description
Integrity in Social Networks and Media 2024, the 5th edition of the INTEGRITY workshop series, is an event co-located within the ACM WSDM’24 conference happening on 4–8 March, 2024 in Mérida, Yucatán (Mexico). The INTEGRITY’24 workshop will be held on March 8th, during the last day of the ACM WSDM conference.
The workshop aims to bring together academic and industry experts working on social media trust and safety to discuss algorithmic and systems aspects of detecting and mitigating integrity challenges and explore opportunities for cross-organizational collaborations. In light of the recent proliferation of generative AI, the themes for this year include integrity risks and opportunities with generative AI and multimodal content understanding for integrity, but we also encourage submissions on broader topics in integrity if they are sufficiently relevant.
List of Topics
- Emerging Threat Vectors: Emerging integrity risks in social media, especially in the context of generative AI, sensitive populations, and elections.
- Detection and Mitigation of Generative AI Powered Risks: Integrity defense against hyper-realistic misinformation, sophisticated evasion techniques, large scale attacks, leveraging behavioral and content anomaly detection, red teaming.
- Foundational Models for Integrity: Generative AI for content moderation, (open sourcing) oracle models for integrity, training datasets.
- Fairness and Responsible AI: Mitigating biases in sensitive attributes such as gender, race, sexual orientation, and political affiliation, overcoming biases in training corpora, preventing discrimination in moderation systems.
- Scaling to Low Resource Languages: Creative approaches for scaling LLMs to low resource languages, performance guarantees, data needs.
- Evaluation Best Practices: Best practices and practical approaches for evaluating risks and measuring progress on integrity issues.
- Addressing Over-enforcement, Subjectivity and Regulatory Needs: Balancing between objective content standards, subjective perceptions of harms and regional regulatory needs.
- Multimodal Harmful Content Understanding: Challenges, opportunities and successful techniques for multi-modal content understanding, dealing with scale for video content - efficient architectures, hardware acceleration, practical heuristics.
- Human In The Loop for Integrity: Quality, efficiency and challenges for human in the loop processes - content labeling, malicious behavioral patterns, emotional impact on human labellers, tooling, LLM assisted labeling.
- Virality and Creator-Audience Dynamics: Detection and mitigation of subscriber-base-growth via low quality, integrity, or genAI produced content, audience expectation and impact on original creators.
Submission Guidelines
We invite two types of submissions, technical papers and talk proposals:
- Technical manuscripts must be 8 pages long for full papers, and 4 pages long for short papers. We invite technical papers of the following types: analysis paper (focus is to generate new insights, rather than the specific method applied), methodology paper (focus is to test the effectiveness of a proposed method), reproduction paper (reproduce results documented in prior work), resource paper (presents a new resource, such as a dataset or tool), and use case paper (presents new insights about a specific use case, such as an event or a community).
- Talk proposals should be 2 pages long, describing the content of a roughly 20-minute talk (the actual length will be determined based on program constraints). We invite submissions from scholars, activists, developers, lawyers, ethics experts, fact-checkers, public servants, journalists, and all-around researchers.
All submissions must be original and not simultaneously submitted to another journal or conference. They must be written in English, and formatted using the standard two-column ACM Sigconf proceedings format. The submission is single-blind.
Accepted papers will either be presented as contributed talks, or as posters. All accepted submissions will be included in the workshop proceedings (CEUR-WS).
Please submit papers and talk proposals through EasyChair: https://easychair.org/conferences/?conf=integrity24
Important Dates
- Paper & Talk proposal submission: 22 Jan 2024
- Notifications: 1 Feb 2024
- Workshop date: 8 March 2024
More information: http://integrity-workshop.org
Organizing Committee
- Lluis Garcia-Pueyo, Meta
- Symeon Papadopoulos, ITI-CERTH
- Panayiotis Tsaparas, University of Ioannina
- Prathyusha Senthil Kumar, Meta
- Aristides Gionis, KTH Royal Institute of Technology
- Vasilis Verroios, Meta
- Giuseppe Manco, ICAR-CNR
- Anton Andryeyev, Meta
- Stefano Cresci, IIT-CNR
- Timos Sellis, Archimedes / Athena Research Center
- Anthony McCosker, Swinburne Social Innovation Research Institute
Contact
All questions about submissions should be emailed to integrity-workshop@googlegroups.com