# 'Crafting Digital Twins' Seminar Series

:::info
**Today's Topic: Knowledge Graphs**
Date & Time: April 25th, 12-4pm, hybrid at Alan Turing Institute, Ada Lovelace Meeting room
Zoom Link: [Here]( https://turing-uk.zoom.us/j/93656786026?pwd=MmoyV0VuR3Y3QUFKNytpL3VDYWh5UT09&from=addon)

👇 Scroll down for Talk titles & abstracts
:::

**What?** [TRIC-DT Seminar Series](https://github.com/alan-turing-institute/tric-dt/tree/main/Seminars) is an interdisciplinary platform for [TRIC-DT](https://www.turing.ac.uk/research/research-projects/tric-dt) researchers & friends to **share and discuss the computational methods, algorithms, and models that underpin Digital Twin technology across diverse fields**.

**Who?** ***Everyone** intersted in digital twin research*

:::info
***If you don't receive email announcements about those seminars (start with [TRIC-DT Seminar Series] in subject matter - get in touch to be added to our mailing list!!!***

---> [e-mail]
:::

## Agenda

| Time | Activity |
| ---- | -------- |
| 12:00 | Welcome & Updates |
| 12:15 | (Guest Speakers:) **Jethro Akroyd & Jiaru Bai** "Beyond Connected Digital Twins: The World Avatar"
|13:00| Lunch break|
|14:00 | (Infrastructure Theme:) **Xiaoxue Shen** "Knowledge graphs for digital twins of structural dynamic systems" & **Ziad Ghauch** "Next-Generation Digital Twins: a Deep Reinforcement Learning Perspective"|
|14:45| (Natural Environment Theme:) **James Byrne** Can Knowledge Graphs Enhance Dynamic Service Integration across polar operations and research? |
|15:30 | General Discussion|

## Guest Speaker

**Dr. Jethro Akroyd & Jiaru Bai**
_Computational Modelling Group
University of Cambridge, Department of Chemical Engineering_
**Beyond Connected Digital Twins: The World Avatar**
The advent of digitalization brings forth new opportunities for data-centric engineering, leading to increased efficiency in the realms of science, engineering, and society. However, the implications of digitalization for environmental and social sustainability remain uncertain. We suggest that knowledge graph technology plays a pivotal role in addressing this challenge. In this presentation, we will introduce “The World Avatar”, a dynamic knowledge graph designed to enable a comprehensive scalable semantic representation of physical and abstract entities, computational agents, and software. We will elucidate the core concepts and principles underlying this idea, and will highlight work to date that shows how these ideas have been and are being applied to real-world problems to enable data-centric decision making and operation to enhance efficiency, reduce costs, avoid emissions, and bolster resilience. We will focus in particular on the contributions that we have made to the ability to represent goals, and how we have applied to contribute to achieving self-driving chemical laboratories.

### Speaker Bios
**Dr Jethro Akroyd**

Jethro is a research-focused chartered chemical engineer who works at the interface of academia and industry, with experience cutting across multinationals, start-ups and universities. He is a Fellow Churchill College and holds positions at CMCL, a multi-award-winning SME, and as a Senior Research Associate in the Computational Modelling Group at the Department of Chemical Engineering and Biotechnology (CEB), University of Cambridge. Jethro has worked extensively with the Cambridge Centre for Advanced Research and Education in Singapore (CARES) to develop solutions to enable the cross-domain interoperability of data and models, with a focus on applications relating to sustainability and the energy transition. He recently led work by CMCL to implement the first cross-sector digital twin of energy, water and telecoms infrastructure networks as part of the Climate Resilience Demonstrator (CReDo), which is a climate change adaptation digital twin demonstrator project to improve resilience across infrastructure networks.

**Dr Jiaru Bai**

Jiaru completed a BEng degree in Chemical Engineering at the University of Manchester with first-class honoursk, winning the Best Final Year Student Award in 2018 He completed an MPhil in Advanced Chemical Engineering at the University of Cambridge in 2019, and a PhD in 2023 in the Computational Modelling Group with a thesis titled "A dynamic knowledge graph approach to self-driving chemical laboratories". In his current research he is developing the use of dynamic knowledge graphs for scalable and reproducible chemical science.

## Meeting Q&A

- []

- []

## Notes pre-meeting Feb 28th
- motivation: motivate why joining data together into a knowledge graph is beneficial, how has CreDo approached interoperability, different domains as examples for where the approach has been used. Jiaru will go into detail on chemical lab automation
- [nature comms paper](https://www.nature.com/articles/s41467-023-44599-9)
- Markus eager to do a short presentation on DT Base to kickstart a brainstorming session
- important Question: what is the question to which KG is a solution?
- ability to support reasoning (ontologies for data organization)
- world as system of interconnected data & systems
- AI lab thinking about optimisation problems
- digital twin data common
- IMFe work with NERC: how can different environmental systems be coupled together? including application of FAIR data principles. ISO19115 standard came up (representation of geospatial datasets), which is not very accessible, can KG graphs help with this, but unclear how exaclty they will help
- Jethro: have looked into this quite a bit. conclusion was to use existing standards, using relational DBs for geospatial data and combining this with KG, two data standards describing slightly different things can be combined. Also include a computational agent that can access the data in different forms and interface with computational services
- [DCE paper](https://www.cambridge.org/core/journals/data-centric-engineering/article/universal-digital-twin-a-dynamic-knowledge-graph/FD25CDFF886CD2ED33D1FDFC13F6BEAB)
- potential projcts for breaktouts: autonomous marine planning team, including planning agents; process models (interactions of different models for planning)
- will we have a debate about definitions around digital twins? will we get derailed?
- James: digital twin is the long-term goal. We want to build interoperable systems first and foremost
- Jethro: we regard DTs as build with ontologies, pragmatic approach to building ontologies, distinct from a string hierarchical model of ontologies (eg national digital twin programme), hence won't give strict definition around knowledge graphs
- Infrastructure talk: beginners lesson in what is a neurosymbolic AI? Third wave of AI?
- DTs will require logical reasoning, which current large AI systems (eg LLMs) are particularly bad at
- Jethro: chemical information chatbots through KG embedding, eg using LLMs to do the query
- jiaru: interest in RL, will the agents be LLM-based?
- Karen booked Florence Nightingale room (1st floor) 1:30-2:30pm for Jethro's meeting with CreDo contact. To be refined once Jethro has arranged meeting. Additionally booked same room 4:00-5:00pm. Teams link included in both if useful.

----

### Link List ✨

#### All Slidedecks
* Guest speaker slides can be [downloaded here](https://www.dropbox.com/scl/fi/n6c54dx0wc2ny64op2ykn/2024_04_25_Turing_TWA_shared.pdf?rlkey=56emkq02pilng58z8pg9h8ut0&e=1&dl=0)
* All other slides will be made available through our [Seminar Series folder](https://thealanturininstitute-my.sharepoint.com/:f:/r/personal/sarana_turing_ac_uk/Documents/TRICDT_Seminar_Series/2024_Apr25_KnowledgeGraphs?csf=1&web=1&e=Tfy0cX)
If you would like to have access please join our Seminar mailing list by emailing us at [e-mail]

#### General links
* [TRIC-DT website](https://www.turing.ac.uk/research/research-projects/tric-dt)
* [Seminar series page](https://github.com/alan-turing-institute/tric-dt/tree/main/Seminars) and [suggested topics](https://github.com/alan-turing-institute/tric-dt/issues?q=is%3Aopen+is%3Aissue+label%3A%22seminar+series%22)
* [TRIC-DT Glossary](https://hackmd.io/npUXO9llSDqagVM3SntR_w)
* [HackMD guide](https://hackmd.io/@turingway/hackmd-guide)

### Code of conduct

* [Take a moment to read this](https://www.turing.ac.uk/events/policies-and-guidelines)

---