S03¶


Research	R0051 — Fact-Checking Gap
Run	2026-03-31
Query	Q001
Search	S03

WebSearch — Computational fact-checking pipeline evidence quality scoring

Summary¶

Field	Value
Source/Database	WebSearch
Query terms	computational fact-checking pipeline evidence quality scoring structured methodology ClaimBuster FAVEL
Filters	None
Results returned	10
Results selected	3
Results rejected	7

Selected Results¶

Result	Title	URL	Rationale
S03-R01	Toward Automated Fact-Checking (ClaimBuster)	https://ranger.uta.edu/~cli/pubs/2017/claimbuster-kdd17-hassan.pdf	Foundational computational fact-checking pipeline — check for evidence quality component
S03-R02	Towards Automated Factchecking: Annotation Schema	https://arxiv.org/pdf/1809.08193	Annotation schema for fact-checking — potential evidence quality structure
S03-R03	DEFAME: Dynamic Evidence-based Fact-checking	https://arxiv.org/html/2412.10510v2	Recent evidence-based pipeline — check for quality scoring

Rejected Results¶

Result	Title	URL	Rationale
S03-R04	Towards Automated Fact-Checking of Real-World Claims	https://ceur-ws.org/Vol-3986/paper2.pdf	Real-world claims focus but no evidence quality framework
S03-R05	The Quest to Automate Fact-Checking (2015)	http://cj2015.brown.columbia.edu/papers/automate-fact-checking.pdf	Early overview — predates current pipeline development
S03-R06	FACT5: A Novel Benchmark	https://aclanthology.org/2025.fever-1.8.pdf	Benchmark for nuanced verdict — task evaluation, not evidence quality
S03-R07	Automated check-worthy sentence detection	https://pmc.ncbi.nlm.nih.gov/articles/PMC9916500/	Claim detection only — upstream of evidence evaluation
S03-R08	ClaimBuster: end-to-end system (ResearchGate)	https://www.researchgate.net/publication/319597830_ClaimBuster_the_first-ever_end-to-end_fact-checking_system	Duplicate of S03-R01 content
S03-R09	Toward Automated Fact-Checking (ACM)	https://dl.acm.org/doi/10.1145/3097983.3098131	Duplicate of S03-R01 (same paper, ACM version)
S03-R10	Towards Automated Factchecking (Semantic Scholar)	https://www.semanticscholar.org/paper/Toward-Automated-Fact-Checking:-Detecting-Factual-Hassan-Arslan/6e99d06b4f2a53f8f6d1f5a51c4fbbee45322ab0	Duplicate of S03-R01

Notes¶

Computational fact-checking pipelines focus on claim detection (check-worthiness scoring), evidence retrieval, and verdict prediction. None of the examined pipelines include structured evidence quality scoring comparable to GRADE. The evidence retrieval step treats all retrieved evidence equally — there is no hierarchical quality assessment of the evidence itself. This is a notable architectural absence.