A dynamic knowledge graph approach to distributed self-driving laboratories

Nature Communications, Journal Year: 2024, Volume and Issue: 15(1)

Published: Jan. 23, 2024

Abstract The ability to integrate resources and share knowledge across organisations empowers scientists expedite the scientific discovery process. This is especially crucial in addressing emerging global challenges that require solutions. In this work, we develop an architecture for distributed self-driving laboratories within World Avatar project, which seeks create all-encompassing digital twin based on a dynamic graph. We employ ontologies capture data material flows design-make-test-analyse cycles, utilising autonomous agents as executable components carry out experimentation workflow. Data provenance recorded ensure its findability, accessibility, interoperability, reusability. demonstrate practical application of our framework by linking two robots Cambridge Singapore collaborative closed-loop optimisation pharmaceutically-relevant aldol condensation reaction real-time. graph autonomously evolves toward scientist’s research goals, with effectively generating Pareto front cost-yield three days.

Language: Английский

---

The advancement of artificial intelligence in biomedical research and health innovation: challenges and opportunities in emerging economies

Globalization and Health, Journal Year: 2024, Volume and Issue: 20(1)

Published: May 21, 2024

The advancement of artificial intelligence (AI), algorithm optimization and high-throughput experiments has enabled scientists to accelerate the discovery new chemicals materials with unprecedented efficiency, resilience precision. Over recent years, so-called autonomous experimentation (AE) systems are featured as key AI innovation enhance research development (R&D). Also known self-driving laboratories or acceleration platforms, AE digital platforms capable running a large number autonomously. Those rapidly impacting biomedical clinical innovation, in areas such drug discovery, nanomedicine, precision oncology, others. As it is expected that will impact healthcare from local global levels, its implications for science technology emerging economies should be examined. By examining increasing relevance contemporary R&D activities, this article aims explore health highlighting implications, challenges opportunities economies. presents an opportunity stakeholders co-produce knowledge landscape health. However, asymmetries capabilities acknowledged since suffers inadequacies discontinuities resources funding. establishment decentralized infrastructures could support overcome restrictions opens venues more culturally diverse, equitable, trustworthy health-related through meaningful partnerships engagement. Collaborations innovators facilitate anticipation fiscal pressures policies, obsolescence infrastructures, ethical regulatory policy lag, other issues present Global South. Also, improving cultural geographical representativeness contributes foster diffusion acceptance worldwide. Institutional preparedness critical enable navigate coming years.

Language: Английский

---

IUPHAR Review: New strategies for medications to treat substance use disorders

Pharmacological Research, Journal Year: 2024, Volume and Issue: 200, P. 107078 - 107078

Published: Jan. 21, 2024

Substance use disorders (SUDs) and drug overdose are a public health emergency safe effective treatments urgently needed. Developing new medications to treat them is expensive, time-consuming, the probability of compound progressing clinical trials obtaining FDA-approval low. The small number FDA-approved for SUDs reflects low interest pharmaceutical companies invest in this area due market forces, characteristics population (e.g., stigma, socio-economic legal disadvantages), high bar regulatory agencies set medication approval. In consequence, most research on funded by government agencies, such as National Institute Drug Abuse (NIDA). Multiple scientific opportunities emerging that can accelerate discovery development SUDs. These include fast efficient tools screen molecules, discover targets, big data explore large sets artificial intelligence (AI) applications make predictions, precision medicine individualize optimize treatments. This review provides general description these strategies with emphasis gaps opportunities. It includes brief overview rising toll SUDs; justification, challenges, develop medications; discussion treatment endpoints being evaluated support from NIDA.

Language: Английский

---

Transforming research laboratories with connected digital twins

Deleted Journal, Journal Year: 2024, Volume and Issue: 1(1), P. 100004 - 100004

Published: Feb. 5, 2024

To substantially expedite scientific discovery, research laboratories need to be further automated. In this regard, the community envisions an 'AI scientist' capable of planning, conducting, and assessing experiments based on higher-order goals reasoning capabilities. We argue that a paradigm shift is necessary bridge gap between current trajectory lab automation vision. Adopting systems perspective reveals several key challenges must addressed. achieving holistic requires network comprehensive distributed digital twins grounded in universal knowledge model. Dynamic graphs are expected play important role, we introduce framework encompassing all aspects experimental research, including infrastructure peripheries. Our considers human-machine interactions from outset empower goal-driven approach brings autonomy.

Language: Английский

---

Leveraging Text-to-Text Pretrained Language Models for Question Answering in Chemistry

ACS Omega, Journal Year: 2024, Volume and Issue: 9(12), P. 13883 - 13896

Published: March 12, 2024

In this study, we present a question answering (QA) system for chemistry, named Marie, with the use of text-to-text pretrained language model to attain accurate data retrieval. The underlying store is "The World Avatar" (TWA), general world consisting knowledge graph that evolves over time. TWA includes information about chemical species such as their and physical properties, applications, classifications. Building upon our previous work on KGQA advanced version Marie leverages fine-tuned Flan-T5 seamlessly translate natural questions into SPARQL queries no separate components entity relation linking. developed QA demonstrates competence in providing results complex involve many hops well showcasing ability balance correctness speed real-world usage. This new approach offers significant advantages prior implementation relied embedding. Specifically, updated boasts high accuracy great flexibility accommodating changes evolution stored without necessitating retraining. Our evaluation underscore efficacy improved system, highlighting its superior compared predecessor.

Language: Английский

---

From Platform to Knowledge Graph: Distributed Self-Driving Laboratories

Research Square (Research Square), Journal Year: 2023, Volume and Issue: unknown

Published: July 25, 2023

Abstract The ability to integrate resources and share knowledge across organisations empowers scientists expedite the scientific discovery process. This is especially crucial in addressing emerging global challenges that require solutions. In this work, we develop an architecture enable distributed self-driving laboratories as part of World Avatar project, which seeks demonstrate how create all-encompassing digital twin based on a dynamic graph. Our approach utilises ontologies capture data material flows involved design-make-test-analyse cycles, employs autonomous agents executable components carry out experimentation workflow. All provenance recorded following FAIR principles, ensuring its accessibility interoperability. We practical application our framework by linking two robots Cambridge Singapore achieve collaborative closed-loop optimisation for pharmaceutically-relevant aldol condensation reaction real time. graph evolves autonomously while progressing towards research goals set scientist. effectively produced Pareto front cost-yield problem over course three days operation.

Language: Английский

---

Marie and BERT─A Knowledge Graph Embedding Based Question Answering System for Chemistry

ACS Omega, Journal Year: 2023, Volume and Issue: 8(36), P. 33039 - 33057

Published: Aug. 25, 2023

This paper presents a novel knowledge graph question answering (KGQA) system for chemistry, which is implemented on hybrid embeddings, aiming to provide fact-oriented information retrieval chemistry-related research and industrial applications. Unlike other existing designs, the operates multiple embedding spaces, use various methods queries spaces in parallel. With answers returned from leverages score alignment model adjust answer scores rerank answers. Further, implements an algorithm derive implicit multihop relations handle complexities of deep ontologies improve answering. The also BERT-based bidirectional entity-linking enhance robustness accuracy module. uses joint numerical efficiently filtering questions. it can invoke semantic agents perform dynamic calculations autonomously. Finally, KGQA handles numerous chemical reaction mechanisms using parsing supported by Linked Data Fragment server. evaluates each module within with chemistry data set.

Language: Английский

---

Where artificial intelligence stands in the development of electrochemical sensors for healthcare applications-A review

TrAC Trends in Analytical Chemistry, Journal Year: 2024, Volume and Issue: 181, P. 117999 - 117999

Published: Oct. 5, 2024

Language: Английский

---

Hacking decarbonization with a community-operated CreatorSpace

Chem, Journal Year: 2024, Volume and Issue: 10(4), P. 1071 - 1083

Published: Jan. 26, 2024

The pressing challenge of decarbonization encompasses a vast combinatorial space interlinked technologies, thus necessitating an increased reliance on artificial intelligence (AI)-assisted molecular modeling and data analytics. Our backcasting analysis proposes future rich in efficient such as sustainable fuels for aviation shipping, well carbon capture utilization. We then retrace the path to this proposed with guidance two constraints: maximization scientists' creative capacities evolution world-centric AI. exploration leads us concept "CreatorSpace," distributed digital system resembling existing hackerspaces makerspaces known accelerating prototyping new technologies worldwide. CreatorSpace serves virtual, semantic platform where chemists, engineers, materials scientists can freely collaborate, integrating chemical knowledge cross-scale, cross-technology tools, operations. This streamlined molecular-to-process-design pathway facilitates diverse array solutions other sustainability technologies.

Language: Английский

---

Research Progresses and Applications of Knowledge Graph Embedding Technique in Chemistry

Journal of Chemical Information and Modeling, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 20, 2024

A knowledge graph (KG) is a technique for modeling entities and their interrelations. Knowledge embedding (KGE) translates these relationships into continuous vector space to facilitate dense efficient representations. In the domain of chemistry, applying KG KGE techniques integrates heterogeneous chemical information coherent user-friendly framework, enhances representation data features, beneficial downstream tasks, such as property prediction. This paper begins with comprehensive review classical contemporary methodologies, including distance-based models, semantic matching neural network-based approaches. We then catalogue primary databases employed in chemistry biochemistry that furnish KGs essential data. Subsequently, we explore latest applications focusing on risk assessment, prediction, drug discovery. Finally, discuss current challenges provide perspective potential future developments.

Language: Английский

---