1st Workshop on Incentives and Contribution Evaluation (WICE 2027)
Reykjavík, Iceland — May 3, 2027
A workshop co-located with SaTML 2027.
Overview
WICE 2027 aims to advance the study of incentives and contribution evaluation in collaborative machine learning. As federated learning moves beyond training a single shared model toward multi-agent systems, and as data valuation, incentive compatibility, and privacy-preserving computation become central to any large-scale collaboration, quantifying each participant’s contribution — fairly, robustly, and without exposing private data — has never been more critical.
This workshop explores contribution evaluation metrics and their axiomatic foundations, the robustness of contribution scores against manipulation and poisoning, incentive mechanisms and mechanism design for collaborative learning, and credit assignment in multi-agent and LLM-agent collaboration, addressing key challenges including compatibility with privacy-enhancing technologies, system and statistical heterogeneity, and verifiable reward distribution. By bringing together researchers and practitioners from the federated learning, security and privacy, data valuation, and game theory communities, WICE fosters discussion on the incentive foundations that determine whether collaborative learning is worth participating in for those involved — co-located with IEEE SaTML 2027 in Reykjavík, Iceland, on May 3, 2027.
Call for Papers
We invite short papers of up to 6 pages excluding references, in the double-column IEEE conference format, submitted in a single track. Suitable submissions include novel preliminary results, systematization and position papers, and condensed versions of mature, unpublished work.
Topics of interest include, but are not limited to:
- Contribution evaluation metrics, axiomatic foundations, and scalable approximations (e.g., Shapley-based measures)
- Contribution evaluation under privacy-enhancing technologies (secure aggregation, differential privacy, trusted hardware)
- Robustness of contribution scores: manipulation, gaming, poisoning, and defenses
- Incentive mechanisms, payments, model-based rewards, and mechanism design for collaborative learning
- Free-riding, misbehavior detection, and accountability
- Fairness of reward and contribution allocation
- Data valuation, data and model marketplaces, and pricing of data and privacy
- Game-theoretic models of participation, coalition formation, and interdependent privacy
- Contribution evaluation under system and statistical heterogeneity: non-IID data, stragglers, intermittent availability, and partial participation
- Cost-aware incentives for the computation, communication, and energy expended by participants, particularly in cross-device, edge, and IoT federations
- Contribution- and incentive-aware client selection, personalization, and federated unlearning
- Verifiable and auditable reward distribution, including blockchain- and smart-contract-based incentive mechanisms
- Contribution evaluation and incentives for collaborative fine-tuning of foundation models and large language models
- Credit assignment, incentives, and contribution evaluation in multi-agent (including LLM-agent) collaboration
- Deployment experiences, benchmarks, and case studies from real-world cross-silo and cross-device federations
Review process
Reviewing is double-blind; each paper will receive at least two reviews.
Publication
Accepted papers will be presented at the workshop and published in the workshop proceedings alongside the IEEE SaTML 2027 conference proceedings. In addition, authors of all submissions will be invited to submit extended versions to an associated journal special issue, subject to the journal’s independent peer review. More info at the bottom of this page.
Submission Site
Papers must be submitted through OpenReview:
https://openreview.net/group?id=IEEE.org/SaTML/2027/Workshop/WICE
OpenReview’s moderation policy for newly created profiles:
- New profiles created without an institutional email will go through a moderation process.
- New profiles created with an institutional email will be activated automatically.
We encourage authors to create or update their OpenReview profiles well before the submission deadline (December 18).
Policies
Submissions must be original and not published or under review elsewhere; the workshop follows the IEEE Code of Conduct
Important Dates
All deadlines are 11:59 PM AoE (Anywhere on Earth).
| Milestone | Date |
|---|---|
| Paper submission | December 18, 2026 |
| Acceptance notification | January 31, 2027 |
| Camera-ready (for proceedings) | March 31, 2027 |
| Workshop | May 3, 2027 |
Organizers
Program Committee
TBA
Contact
pejo@crysys.hu
Program
TBA
Keynotes
TBA
Accepted papers
TBA
Journal Special Issue
WICE partners with Frontiers in Computer Science (Computer Security section) on a companion Research Topic, Incentives and Contribution Evaluation in Secure Collaborative Learning, edited by the workshop co-chairs: Balázs Pejó, Steve Drew, Qiang Tang, and Delio Jaramillo-Velez.
The Research Topic treats contribution evaluation as both a security mechanism and an attack surface in federated learning. It welcomes original research, systematization-of-knowledge, survey, and perspective articles on topics such as attacks on and defenses for contribution scoring, privacy-preserving and Byzantine-robust evaluation, secure and verifiable reward distribution, mechanism design, data valuation, and credit assignment in multi-agent and LLM-agent collaboration.
Authors and attendees of WICE are warmly invited to submit substantially extended versions, and the Research Topic is also open to original submissions from outside the workshop.