Latent Dynamics · 2026-11-04

A research world model of power engineering learns its dynamics and reveals what-ifs

Fanping Sui, Ying Zhang, Liwei Lin

arXiv:2611.04566PDF

AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.

TL;DR

The authors introduce “research world models,” treating a research field as a dynamical system with interventions, and build one for power engineering from 1.1M papers. Typed concept hypergraphs are read from abstracts; two graph-temporal cores learn concept growth and first-time method–target links; intervention operators estimate what-if effects from natural experiments. In temporally held-out tests they forecast years ahead and match held-out intervention effects, and they answer natural-language what-ifs as paired simulations.

Why it matters

The paper targets research planning in path-dependent fields like power engineering, where counterfactual consequences of interventions are not directly observable. By learning a transition for a research field and estimating intervention effects, the model aims to simulate and test research decisions before they are made, and it reports validation against held-out natural experiments and concept/link forecasting.

Method

  • Reads 1.20M power-engineering abstracts into role-typed concepts and a monthly typed hypergraph via a distilled language-model reader.
  • Learns dynamics with two temporal graph-network cores: a growth core for log growth forecasts (12–60 months) and a link core for which concept pairs will be linked for the first time.
  • Learns intervention operators from natural experiments to estimate effects of actions; answers what-if questions by paired simulations with and without the intervention, audited by an independent judge model.
Abstract (from arXiv)

Decisions about the direction of research in power engineering determine which technologies mature and, through them, shape infrastructure that lasts decades, yet their consequences cannot be observed before they are taken, and existing forecasts of research fields extrapolate a single trajectory. Here we introduce research world models, which treat a research field as a dynamical system that can be acted upon, and establish one for power engineering from 1.1 million papers. A distilled language model reads each paper into role-typed concepts, two sophisticated, high-fidelity graph-temporal neural network cores learn how the resulting concept network evolves, and operators learned from natural experiments in the publication record estimate the effects of interventions. In temporally held-out tests, the model forecast years ahead which concepts would grow and which method-target pairs would be joined for the first time, outperforming tuned machine-learning baselines in most settings, and its estimated intervention effects tracked those realized in natural experiments it had not seen. Its training scheme and three of its operators replicated in two fields withheld from development. Questions posed in natural language are compiled by a domain-specialized 27-billion-parameter language model into executable actions and answered as paired simulations with and without the intervention. The answers were rated 4.1 of 5 for faithfulness by an independent language-model judge. By revealing the latent dynamics of a research field and its response to interventions, research world models offer a new way to simulate and test research decisions before they are made.

Related papers

Publication record

Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.

First announced by arXiv: 2026-11-04 · Indexed by Pub-AI: 2026-11-04 UTC

preprint

A research world model of power engineering learns its dynamics and reveals what-ifs

Fanping Sui, Ying Zhang, Liwei Lin

First announced by arXiv: 2026-11-04 · Original record ↗

arxiv
2611.04566
Explicit source versions (1)
  • v1 · Date not supplied
  1. Preprint available

    Source date: 2026-11-04 · First detected: 2026-11-04 UTC · Historical baseline

    Source evidence (1)
    • Official evidence ↗ · RSS guid=oai:arXiv.org:2611.04566v1; arxiv:announce_type=new; item pubDate (official metadata)
Abstract · arxiv

Decisions about the direction of research in power engineering determine which technologies mature and, through them, shape infrastructure that lasts decades, yet their consequences cannot be observed before they are taken, and existing forecasts of research fields extrapolate a single trajectory. Here we introduce research world models, which treat a research field as a dynamical system that can be acted upon, and establish one for power engineering from 1.1 million papers. A distilled language model reads each paper into role-typed concepts, two sophisticated, high-fidelity graph-temporal neural network cores learn how the resulting concept network evolves, and operators learned from natural experiments in the publication record estimate the effects of interventions. In temporally held-out tests, the model forecast years ahead which concepts would grow and which method-target pairs would be joined for the first time, outperforming tuned machine-learning baselines in most settings, and its estimated intervention effects tracked those realized in natural experiments it had not seen. Its training scheme and three of its operators replicated in two fields withheld from development. Questions posed in natural language are compiled by a domain-specialized 27-billion-parameter language model into executable actions and answered as paired simulations with and without the intervention. The answers were rated 4.1 of 5 for faithfulness by an independent language-model judge. By revealing the latent dynamics of a research field and its response to interventions, research world models offer a new way to simulate and test research decisions before they are made.