PROMETHEUS: Automating Deep Causal Research Integrating Text, Data and Models

2026-05-12

AI Summary

This work addresses the challenge of integrating fragmented causal claims scattered across texts, datasets, and models into a navigable, verifiable, and evidence-grounded global causal knowledge system. It proposes a “causal atlases” framework grounded in topos-theoretic world models and sheaf-theoretic structures, organizing heterogeneous sources into a covering family of local causal predictive state models. Through restriction maps and gluing diagnostics, the framework enables rigorous analysis of consistency, contradictions, and uncertainties. The approach supports explicit claim localization, evidential traceability, and counterfactual validation, thereby overcoming the limitations of conventional monolithic causal graphs. Empirical demonstrations—including assessments of ocean temperature impacts, GLP-1–mediated weight loss efficacy, and health claims about resveratrol—show successful counterfactual evaluation and model reconstruction directly from raw data and code.

Abstract

Large language models can extract local causal claims from text, but those claims become more useful when organized as persistent, navigable world models rather than as flat summaries. We introduce PROMETHEUS, a framework that turns retrieved literature, filings, reviews, reports, agent traces, source data, code, simulations, and scientific models into causal atlases: sheaf-like families of local causal predictive-state models over an explicit cover of a research substrate. Each local region contains causal episodes, structured claim tables, predictive tests, support statistics, and provenance; restriction maps compare overlapping regions; gluing diagnostics expose agreement, drift, contradiction, and underdetermination. The resulting Topos World Model is not a single universal graph. It is a research instrument for navigating what a corpus says, where it says it, how strongly it is supported, and where local claims fail to assemble into a coherent global view. Three literature-atlas case studies -- ocean-temperature impacts on marine populations, GLP-1 weight-loss evidence, and resveratrol/red-wine health-benefit claims -- illustrate deep causal research from text with explicit locality, evidence, persistent state, and gluing tension. Four grounded-counterfactual case studies -- a Nature Climate Change microplastics forcing paper, an Indus Valley hydrology paper with VIC-derived figure data and model code, the canonical Sachs protein-signaling study with single-cell perturbation data, and a Nature singing-mouse study with MAPseq projection matrices -- show a stronger mode: when a paper ships source data, simulation outputs, or code, PROMETHEUS can evaluate a counterfactual against that scientific substrate and then rebuild the sheaf world model around the

Problem

• causal modeling
• knowledge integration
• scientific literature
• counterfactual reasoning
• world models

Innovation

• causal atlas
• sheaf theory
• grounded counterfactuals
• Topos World Model
• predictive-state models