Urban Vulnerability Assessment of Sea Level Rise in Singapore through the World Avatar DOI Creative Commons
Shin Zert Phua, Kok Foong Lee,

Yi-Kai Tsai

et al.

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(17), P. 7815 - 7815

Published: Sept. 3, 2024

This paper explores the application of The World Avatar (TWA) dynamic knowledge graph to connect isolated data and assess impact rising sea levels in Singapore. Current level rise vulnerability assessment tools are often regional, narrow scope (e.g., economic or cultural aspects only), inadequate representing complex non-geospatial consistently. We apply TWA conduct a multi-perspective Singapore, evaluating vulnerable buildings, road networks, land plots, sites, populations. introduce OntoSeaLevel, an ontology describe scenarios, its on broader elements defined other ontologies such as buildings (OntoBuiltEnv ontology), networks (OpenStreetMap plots (Ontoplot Ontozoning ontology). deploy computational agents synthesise from government, industry, publicly accessible sources, enriching with metadata property usage, estimated construction cost, number floors, gross floor area. An agent is applied identify instantiate impacted sites using OntoSeaLevel. These include populations at risk. showcase these unified visualisation, demonstrating TWA’s potential planning tool against through assessment, resource allocation, integrated spatial planning.

Language: Английский

Cross-Domain Knowledge Transfer without Retraining to Facilitating Seamless Knowledge Application in Large Language Models DOI Creative Commons
Jae Hoon Kim,

Hye Rin Kim

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: April 29, 2024

Abstract Cross-domain knowledge transfer in large language models (LLMs) presents significant challenges, particularly regarding the extensive resources required for retraining. This research introduces innovative embedding adaptation and context adjustment techniques that enable LLMs to efficiently across diverse domains without need comprehensive Experimental results demonstrate improved model flexibility reduced computational demands, highlighting potential rapid deployment scalability. These findings suggest a sustainable approach deploying adaptive AI various sectors, significantly impacting future developments artificial intelligence.

Language: Английский

Citations

12
Ai-enabled language models (LMs) to large language models (LLMs) and multimodal large language models (MLLMs) in drug discovery and development DOI Creative Commons
Chiranjib Chakraborty, Manojit Bhattacharya, Soumen Pal

et al.

Journal of Advanced Research, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 1, 2025

Language: Английский

Citations

0
Natural language access point to digital metal–organic polyhedra chemistry in The World Avatar DOI Creative Commons
Simon D. Rihm, Dan Tran, Aleksandar Kondinski

et al.

Data-Centric Engineering, Journal Year: 2025, Volume and Issue: 6

Published: Jan. 1, 2025

Abstract Metal–organic polyhedra (MOPs) are discrete, porous metal–organic assemblies known for their wide-ranging applications in separation, drug delivery, and catalysis. As part of The World Avatar (TWA) project—a universal interoperable knowledge model—we have previously systematized MOPs expanded the explorable MOP space with novel targets. Although these data available via a complex query language, more user-friendly interface is desirable to enhance accessibility. To address similar challenge other chemistry domains, natural language question-answering system “Marie” has been developed; however, its scalability limited due reliance on supervised fine-tuning, which hinders adaptability new domains. In this article, we introduce an enhanced database first-of-its-kind tailored chemistry. By augmenting TWA’s geometry data, enable visualization not just empirically verified structures but also machine-predicted ones. addition, renovated Marie’s semantic parser adopt in-context few-shot learning, allowing seamless interaction extensive repository. These advancements significantly improve accessibility versatility TWA, marking important step toward accelerating automating development reticular materials aid digital assistants.

Language: Английский

Citations

0
Question-answering system for combustion kinetics DOI Creative Commons
Laura Pascazio, Dan Tran, Simon D. Rihm

et al.

Proceedings of the Combustion Institute, Journal Year: 2024, Volume and Issue: 40(1-4), P. 105428 - 105428

Published: Jan. 1, 2024

In this paper, we introduce for the first time a natural language question-answering (QA) system specifically designed field of combustion kinetics. This marks significant step towards achieving PrIMe vision as outlined by Frenklach in 2007, offering user-friendly interface that allows researchers and practitioners to easily access query information about chemical mechanisms. QA is key component "The World Avatar" (TWA), dynamic framework built upon semantic web technologies. TWA characterized its layered structure, which includes knowledge graph (KG), software agents, real-world data integration. These layers collectively create comprehensive unified managing analyzing complex from various domains. We detail enhancements made TWA's ontologies (OntoSpecies, OntoKin, OntoCompChem) meet specific challenges kinetics improve their representation accuracy. By focusing on provenance interoperability, our approach ensures transparent reliable management adheres FAIR principles, vital precise retrieval analysis. The role agents populating these highlighted, showcasing how they transform raw into meaningful structured generate new insights within ecosystem. Additionally, technologies' interoperability feature facilitates integration exchange across different platforms tools, making machine-actionable. instantiated KG four H2/O2 five CH4/O2 reaction mechanisms taken literature then demonstrate system's capabilities answering questions related proof concept. Lastly, discuss future directions framework, include not only extensions but also external tool automate tasks such generation kinetic mechanism, further expanding functionality application

Language: Английский

Citations

2
Dynamic control of district heating networks with integrated emission modelling: A dynamic knowledge graph approach DOI Creative Commons

Wanni Xie,

Feroz Farazi,

John Atherton

et al.

Energy and AI, Journal Year: 2024, Volume and Issue: 17, P. 100359 - 100359

Published: May 20, 2024

This paper presents a knowledge graph-based approach for the dynamic control of district heating network with integrated emission dispersion modelling. We propose an interoperable and extensible implementation to forecast anticipated heat demand municipal network, minimise associated total generation cost based on set available sources, couple it modelling corresponding emissions provide automatic insights into air quality implications various sourcing strategies. achieve cross-domain in nexus energy via developed ontologies autonomous software agents, which can be chained together World Avatar graph resemble behaviour complex systems. Furthermore, we have City Energy Analyst this ecosystem building-level potential foster strategic analyses scenario planning. The calculations use building weather data from place inherent assumptions official release, facilitating more data-driven approach. All cases are implemented mid-size town Germany as proof-of-concept, unified visualisation interface is provided, allowing examination 3D buildings alongside their supply time series, well data. With work, outline Semantic Web technologies connect digital twins holistic smart cities, thereby addressing increasing complexity interconnected

Language: Английский

Citations

1
Urban Vulnerability Assessment of Sea Level Rise in Singapore through the World Avatar DOI Creative Commons
Shin Zert Phua, Kok Foong Lee,

Yi-Kai Tsai

et al.

Applied Sciences, Journal Year: 2024, Volume and Issue: 14(17), P. 7815 - 7815

Published: Sept. 3, 2024

This paper explores the application of The World Avatar (TWA) dynamic knowledge graph to connect isolated data and assess impact rising sea levels in Singapore. Current level rise vulnerability assessment tools are often regional, narrow scope (e.g., economic or cultural aspects only), inadequate representing complex non-geospatial consistently. We apply TWA conduct a multi-perspective Singapore, evaluating vulnerable buildings, road networks, land plots, sites, populations. introduce OntoSeaLevel, an ontology describe scenarios, its on broader elements defined other ontologies such as buildings (OntoBuiltEnv ontology), networks (OpenStreetMap plots (Ontoplot Ontozoning ontology). deploy computational agents synthesise from government, industry, publicly accessible sources, enriching with metadata property usage, estimated construction cost, number floors, gross floor area. An agent is applied identify instantiate impacted sites using OntoSeaLevel. These include populations at risk. showcase these unified visualisation, demonstrating TWA’s potential planning tool against through assessment, resource allocation, integrated spatial planning.

Language: Английский

Citations

0