First Call for Papers
Retrieval Beyond Documents (RBD)
The First Workshop on Capability Retrieval for Agents
Co-located with NAACL 2027 in San Francisco, California, USA
Website: https://rbd-workshop.github.io/
RBD brings together the information retrieval, NLP, and agentic AI communities to study how agents retrieve and use tools, skills, workflows, memories, models, and other agents. We welcome research on finding useful resources, combining them effectively, and evaluating their contribution to task completion.
Contributions may include methods and architectures, systems, position papers, benchmarks and datasets, demonstrations, reproducibility studies, and diagnostic or negative results.
Topics include, but are not limited to:
- Resource representation and indexing: description enrichment, metadata, usage examples, execution traces, and evolving indexes.
- Retrieval, ranking, and composition: retrieval methods, reranking, joint selection, resource compatibility, and cost and reliability constraints.
- Retrieval during execution: query reformulation, memory and workflow reuse, and adaptation through observations and feedback.
- Evaluation through use: benchmarks, task completion, error analysis, generalization, and robustness.
- Registries and deployment: dynamic resource collections, MCP ecosystems, permissions, provenance, freshness, and trustworthy resource descriptions.
Submissions
- Direct submissions: anonymized papers for double blind review through the RBD submission portal.
- ARR commitments: papers with completed ARR reviews and a meta-review through the RBD commitment portal.
Please use the ACL template and anonymize your paper and supplementary materials. All authors need an active OpenReview profile; register early if you do not already have one.
Full author guidelines: https://rbd-workshop.github.io/cfp/
Important dates
All deadlines are at 11:59 p.m. Anywhere on Earth (AoE, UTC-12).
- Direct paper submission: February 5, 2027
- ARR commitment: March 12, 2027
- Author notification: March 26, 2027
- Camera ready papers: April 16, 2027
- NAACL 2027: June 1 to 5, 2027
- RBD workshop day: to be announced
Organizers
- Hitesh Laxmichand Patel, Oracle AI
- Akari Asai, Carnegie Mellon University
- Omar Khattab, Massachusetts Institute of Technology
- Danqi Chen, Princeton University
- Zifei Wang, Tencent IMA Product Center
- Avi Sil, Oracle
- Sewon Min, University of California, Berkeley
- Connor Shorten, Weaviate
Contact: rbd-workshop [at] googlegroups.com