MetaSynDec: Agentic Harness for Executable Analytical Knowledge in Meta-Analysis
A new paper on arXiv (2608.01711) introduces the Executable Analytical Knowledge Representation (EAKR), a machine-actionable format for meta-analysis synthesis. The authors argue that structured evidence alone is insufficient for executable computation; decisions on evidence assignment, contrasts, outcome alignment, effect-size formulation, and methodological admissibility must be explicit. EAKR captures evidence, relations, numerical inputs, constraints, provenance, and unresolved issues. The framework is operationalized in MetaSynDec, an agentic harness designed to transform structured evidence into executable meta-analysis. The paper addresses a gap in current automated approaches, which often embed analytical decisions in model outputs or code, lacking independent verifiability. The work is relevant to knowledge-based scientific analysis and computational reproducibility.
Key facts
- Paper arXiv:2608.01711 introduces EAKR (Executable Analytical Knowledge Representation).
- EAKR is a machine-actionable representation for meta-analysis synthesis.
- It includes evidence, relations, numerical inputs, constraints, provenance, and unresolved issues.
- MetaSynDec is an agentic harness operationalizing EAKR.
- The paper argues structured evidence alone is insufficient for executable computation.
- Decisions on evidence assignment, contrasts, outcome alignment, effect-size formulation, and methodological admissibility must be explicit.
- Existing automated approaches embed decisions in model outputs, code, or workflow traces.
- EAKR aims to make analytical knowledge independently verifiable.
Entities
Institutions
- arXiv