Scope
Synthetic data enters the web research stack in two ways. It is deliberately generated to replace, augment, or stress-test scarce and sensitive datasets, and it is increasingly encountered as machine-generated content on the open web. Data used by search engines, recommenders, knowledge systems, moderation pipelines, and web agents may no longer have a simple human-authored provenance chain. The question is not only how to generate realistic samples, but when synthetic data is traceable, useful, representative, and safe to reuse.
SynWeb invites work on any stage of the verification-first lifecycle: generate, constrain, verify, curate, evaluate on web tasks, and release or deploy.
Topics
Topics include, but are not limited to:
- Synthetic web corpus and interaction-data generation
- Multilingual, multimodal, and structured synthetic data
- Retrieval-, knowledge-, schema-, and rule-constrained generation
- Factuality, consistency, uncertainty, and provenance verification
- Detection and measurement of synthetic-content contamination and repeated re-ingestion
- Privacy, licensing, data governance, and reproducible release practices
- Benchmarks that connect intrinsic quality to search, recommendation, fact verification, knowledge extraction, moderation, and agent performance
- Fairness, coverage, and distribution-shift analysis
- Negative results, failure analyses, and practical tooling for auditing synthetic-data pipelines
Submission types
| Type | Length |
|---|---|
| Research paper | Up to 8 pages |
| Short or position paper | Up to 4 pages |
| Dataset, benchmark, tool, demo, or negative result | Up to 4 pages |
Page limits exclude references.
Formatting
Submissions must be in English and use the ACM double-column format (sigconf) from the ACM proceedings template, as required for TheWebConf 2027 workshops. Submit a single PDF.
Review
All submissions undergo double-blind peer review. Remove author names and affiliations, and refer to your own prior work in the third person. Papers are evaluated on relevance to the workshop, technical quality, novelty, and clarity of reporting, including how synthetic data was generated and verified.
Proceedings and non-archival option
Accepted papers will be included in the TheWebConf 2027 Companion Proceedings when eligible under conference policy. Previously published or concurrently reviewed work may be submitted for non-archival presentation only; indicate this at submission time.
At least one author of each accepted paper must register for the workshop and present the work. Accepted contributions are presented as full talks, short talks, lightning talks, or demonstrations depending on type and program space.
How to submit
Questions about scope or eligibility can be sent to sabdaljalil@tamu.edu.