KISS - Knowledge Infrastructure for Scientific Simulation: A Scaffolding for Agentic Earth Science

2026-05-18 · Source: Takara TLDR - Daily AI Papers · Field: Science & Research — Environmental Science & Earth Systems, Artificial Intelligence & Machine Learning, Mathematics & Computational Sciences · Depth: Expert, medium

Summary

KISS (Knowledge Infrastructure for Scientific Simulation) is an agent-actionable scaffold designed to democratize access to complex process-based simulation models in Earth sciences. This infrastructure externalizes scientific expertise into validated modeling operators, staged domain protocols, and diagnostic recovery mechanisms. In a 3,000-trial coupled-hydrology benchmark, agents equipped with KISS achieved physically plausible, verifiable end-to-end simulations in up to 84% of trials, significantly outperforming agents without KISS, which plateaued below 40%. The researchers also developed a Knowledge Dissection Toolkit (KDT) that autonomously constructed 119 KIs, enabling end-to-end agent execution for 117 additional process-based models across 14 Earth-science domains. This demonstrates that operational expertise is structured and extractable, not ad hoc, and lowers access barriers for non-specialists and integration barriers between modeling communities.

Key takeaway

For AI Architects and Research Scientists developing agentic systems for scientific discovery, integrating Knowledge Infrastructure (KI) can dramatically improve simulation reliability and accessibility. Your teams should consider using tools like the Knowledge Dissection Toolkit (KDT) to systematically extract and operationalize domain expertise, thereby enabling agents to perform complex, verifiable simulations across diverse scientific disciplines with higher success rates.

Key insights

Knowledge infrastructure significantly enhances agent performance and accessibility for complex scientific simulations.

Principles

• Operational expertise is structured and extractable.
• Externalizing expertise improves simulation reliability.

Method

The Knowledge Dissection Toolkit (KDT) autonomously constructs Knowledge Infrastructure (KI) by dissecting scientific understanding into validated modeling operators, domain protocols, and diagnostic recovery mechanisms.

In practice

• Use KI to improve agent reliability in simulations.
• Apply KDT to automate KI construction for new models.

Topics

• Knowledge Infrastructure
• Agentic Earth Science
• Process-based Simulation
• Knowledge Dissection Toolkit
• Coupled Hydrology

Code references

• AsiriweLab/S5-HES-Agent

Best for: AI Scientist, Research Scientist, AI Architect

Related on AIssential

• Science Earth: Towards A Planet-Scale Operating System for AI-Native Scientific Discovery — A planet-scale runtime enables AI capabilities to self-organize for distributed, self-correcting scientific discovery.
• K-Dense-AI / scientific-agent-skills — Scientific Agent Skills offers 135 tools for AI agents to perform complex scientific research across diverse domains.
• The uncritical adoption of AI in science is alarming — we urgently need guard rails — Uncritical AI adoption in science risks reduced research quality, narrowed scope, and deskilling future scientists.
• TerraBench: Can Agents Reason Over Heterogeneous Earth-System Data? — Current LLM agents struggle with precise numerical grounding and complex workflow orchestration in Earth-science tasks.
• SkillWiki: A Living Knowledge Infrastructure for Agent Skills — SkillWiki provides a unified infrastructure for agent skill management, from ingestion to evolution, linking skills to evidence.
• Agents-K1: Towards Agent-native Knowledge Orchestration — Agent-native knowledge orchestration transforms raw scientific papers into structured, multimodal knowledge graphs for enhanced LLM reasoning.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Takara TLDR - Daily AI Papers.