Johnson Matthey: Data Science@Work Seminar

In this presentation, Dr Aakash Varambhia showed us the importance of programming skillsets in the chemical engineering space. Students learnt about the process of research in an industrial setting. As opposed to traditional academia, where research can be slow moving and take longer to show up in real world applications, findings in this industry are quickly implemented on the production line.

 

The utility of data science tools such as image analysis, which many Compass students are currently working on, were discussed, with examples of how Johnson Matthey deploys these tools to test and improve its products. It was fascinating to learn about a new industry and where data scientists fit into it.

 

Rachel Wood, Compass CDT student, Cohort 4

 

Title: ‘Delivering State-of-the-Art Imaging Data Science: Catalysis, Characterisation and Beyond’

Speaker: Dr Aakash Varambhia, Data Science Lead at Johnson Matthey

 

Abstract:

Over the past few decades, the pursuit of cleaner energy, sustainable chemicals, and greener mobility solutions has reshaped industrial research and development. At Johnson Matthey (JM), this challenge has brought together a rich mix of materials scientists, chemists, engineers, computer scientists, and mathematicians, all working in concert. At the heart of this transformation is the ability to harness and interpret increasingly complex datasets, ranging from high-resolution microscopy images and detailed spectroscopic data, to correlative 3D tomographic reconstructions.

 

Central to JM’s approach is the integration of multi-lengthscale characterisation methods such as X-ray tomography (XRT), focused ion beam (FIB) tomography, and transmission electron microscopy (TEM) to capture hierarchical structures at varying length scales. This comprehensive pipeline reveals subtle correlations between structure and catalytic performance, from micron-level features uncovered by XRT, through meso- and nano-scale insights offered by FIB, down to atomic-level details exposed by TEM.

 

To handle the resulting volume and variety of data, JM has developed a flexible, scalable platform leveraging Python-based frameworks, parallel computing libraries, and cloud-based infrastructure. This ensures seamless workflows for data ingestion, preprocessing, analysis, and visualization, enabling swift collaboration, informed decision-making, and predictive modelling that iteratively refines both theory and experiment.

 

In this talk, Dr Varambhia showcased JM’s data-driven research in catalysis science, highlighting how the interplay of advanced instrumentation, computational analytics and robust software ecosystems enables new levels of understanding.

 

About the Speaker:

Dr Aakash Varambhia leads the Advanced Characterisation Data Science team at Johnson Matthey’s Technology Centre. He and his colleagues Tom Ellaby, Zaeem Najeeb, and Dogan Ozkaya have pioneered the development of Python-based research platforms and analysis workflows. Their collective efforts ensure that scientists and engineers across JM have access to powerful, scalable tools for handling complex datasets. Through this approach, data science ceases to be a siloed function and instead becomes a core driver of innovation, complementing traditional experimental and theoretical techniques.

 

About Johnson Matthey:

With a distinguished history spanning more than two centuries, Johnson Matthey stands at the forefront of sustainable technologies, applying cutting-edge science to address pressing global challenges. The company’s solutions underpin a broad range of industries from clean energy and chemicals, to automotive applications enabling their clients, partners, and wider society to thrive. By integrating mathematical methodologies and advanced data interrogation with domain-specific knowledge, JM continually pushes the boundaries of what can be achieved, ensuring that research not only meets today’s demands, but also paves the way for a more sustainable tomorrow.

 

Schedule:

  • 2.00pm – 3.00pm – Seminar + Q&A – G.07 (Heilbronn Room), Fry Building
  • 3.00pm – 4.00pm – Refreshments and networking session – Common Room, Fry Building

Advai: DataScience@work seminar

Image of speaker, Damian Ruck from AdvaiThe rise of large language models (LLMs) and artificial intelligence (AI) has led to increasing concerns over the safety and robustness of such models. This presentation by Dr Damian Ruck, Chief Researcher at Advai, highlighted the critical work that AI safety companies are doing to ensure the safe and responsible deployment of AI tools.

As many Compass students and researchers work in the field of generative modelling, this presentation served as an enriching discussion, bridging the gap between theoretical research developments and practical applied concerns. Dr Ruck’s experience and insights into the challenges and future directions of AI safety were particularly valuable.

The rapid advancement of AI underscores the importance of having experts like Dr Ruck share their perspectives on the field. This presentation provided a unique opportunity for our research community to engage with the critical issues surrounding AI safety and security, and to consider their implications for our own research efforts and direction.

Title: Introducing Advai: Advancing AI Safety and Security

Speaker: Damian Ruck, Chief Researcher, Advai 

www.advai.co.uk

Advai, a leading company in the field of Safe and Secure AI, specialises in testing, evaluation, and assistance of AI systems. It conducts research for a range of organizations, with the sole purpose of identifying AI failure modes and ensuring robust performance. Advai collaborates with prominent AI companies, financial services, insurance providers, the National Cyber Security Centre (NCSC), the AI Task Force, and venture capitalists.

Advai’s research team plays a crucial role in the company’s success by providing expertise and original research to support consultancy work. The team develops tools and metrics for Advai’s products and undertakes research contracts for the MoD and other organizations. Their work revolves around three broad areas: adversarial red teaming; tools for testing, evaluation and assurance; and tools for building more robust AI systems.

Looking ahead, Advai’s research team is exploring innovative solutions such as assurance agents, which involve the use of AI agents to semi-automate testing and evaluation processes. They are also investigating the adaptation of traditional system safety processes to AI systems, moving towards formal verification of AI systems, ensuring their reliability and security.

By addressing the critical challenges in AI safety and robustness, Advai is paving the way for the development and deployment of trustworthy AI systems across various industries. Their ongoing research and collaborations with key stakeholders, position Advai at the forefront of shaping the future of Safe and Secure AI.

Damian Ruck attained a PhD at the University of Bristol, before spending several years in U.S. universities publishing Computational Social Science research in top academic journals. He has since taken up the role of Chief Researcher at Advai, an Adversarial Machine Learning startup, applying Data Science skills to building Robust and Secure AI.

Schedule:

  • 2.00 – 3.00pm Introducing Advai: Advancing AI Safety and Security
  • 3.00 – 4.00pm Networking session

Location: Room 2.04, Fry Building

Microsoft Research: DataScience@work seminar

Title: TBA

Speaker: James Hensman, Principal Machine Learning Researcher, Microsoft Research

 

Schedule:

  • 2.00 – 3.00pm Seminar + Q&A (Rm. 2.04, Fry Building)
  • 3.00 – 4.00pm Networking session (Common Room, Fry Building)

Location: Room 2.04, Fry Building 

 

Abstract: TBA

 

 

Kew Gardens: DataScience@work seminar

The presentations highlighted some important and varied research that data scientists can be involved in, such as: identifying plants with potential to curing malaria, efforts to auto-catalogue plants based on their important medical characteristics, and identifying coffee plant species robust to climate change. Discussions between the students and the presenters were enlightening as to the challenges of being a data scientist, particularly in balancing their work time between different projects and funding applications. Overall, it was great to meet the guys from Kew and feedback from all involved was very positive!

  • Daniel Milner, Compass CDT Cohort 3

 

Title: Harnessing Botanical and Mycological Knowledge: Data-Driven Solutions for Global Challenges 

Speakers: Adam Richard-Bollans, Research Fellow, and Eren Karabey,  Kew Gardens

 

 

 

Abstract: 

Harnessing Botanical and Mycological Knowledge: Data-Driven Solutions for Global Challenges 

A vast amount of botanical and mycological knowledge exists within Kew as well as in global datasets and published scientific literature. In this talk Adam and Eren will discuss their work on turning knowledge into data and the ways they are using data to tackle a variety of global challenges; including extracting structured data from scientific papers, analysing phytochemical diversity across the plant tree of life, searching for new medicinal plant-derived compounds, analysing the chemical composition of coffee species and predicting germination rates of orchid seeds. 

 

Speakers:  

Dr. Adam Richard-Bollans is a Future Leader Fellow at RBG, Kew. With a background in mathematics and computing, his main work is focused on machine learning methods to efficiently search the estimated 343,000 known vascular plant species for new compounds of pharmaceutical interest. 

Eren Karabey is a student on placement at Kew as an undergraduate intern. Eren is studying Computer Science and Artificial Intelligence at the University of Bath. His main project is to automatically build machine learning-ready datasets from open access research papers.

About Kew Science 

Scientists at Kew work collaboratively and globally to understand and protect biodiversity and to discover sustainable solutions to some of our biggest global challenges. Kew Science encompasses a wide range of disciplines and activities, from analysing evolutionary processes and ecological interactions, to exploring bioactive molecules to unlock useful properties in plants and fungi.  Kew also holds a globally unique, substantial and growing collection of fungal and plant specimens, illustrations, databases, scientific literature, and archives of unpublished material. Ongoing digitisation of these collections is beginning to unleash new opportunities for large-scale data mining and analysis for scientific discovery and innovation.

 

Schedule:

  • 2.00 – 3.00pm Seminar + Q&A (Rm. 2.04, Fry Building)
  • 3.00 – 4.00pm Networking session (Common Room, Fry Building)

     Location: Room 2.04, Fry Building 

 

David Greenwood: DataScience@work seminar

Dr David Greenwood, mentor to aspiring Data Leaders, gave an inspiring talk on how to approach the industry job market. He shared insights from his diverse, eventful and successful career and some of the valuable lessons he learned during his years in industry. He advised which skills are highly requested in data science and gave valuable suggestions on what to focus on in your first job!

  • Emma Ceccherini, Compass CDT Cohort 4 student

 

Title: Preparing for Data Science at Work

Speaker: David Greenwood, Founder,

Abstract:
Congratulations you’ve got your first job in industry! (Or maybe you want one?) What do you do to prepare? How do you go about preparing? What do you need to do to impress? How do you not be nightmare for your manager!
Who am I? Dr David Greenwood. What do I know about this? That’s a fair question – your time is valuable! I graduated from Bristol with an EngD and subsequently worked at the boundary of academia & industry for a decade plus hiring and mentoring MSc & PhD graduates. I rose up the ranks from Research Engineer to Head of Data Science (and beyond) with a career spanning Engineering, Financial markets and Impact Investment. I now mentor aspiring Data Leaders and you’re stuck with me for about 1 hour – so let’s make our time together count! It’ll be an interactive session so you’ll be doing as much talking as me. We’ll go on a light hearted and hopefully useful exploration on preparing for Data science at work. Bring questions!

Schedule:

  • 2.00 – 3.00pm Seminar + Q&A (Rm. 2.04, Fry Building)
  • 3.00 – 4.00pm Networking session (Common Room, Fry Building)

Location: Room 2.04, Fry Building 

 

 

ONS: DataScience@work seminar

Dr. Tim Green, a data scientist at the Office for National Statistics Data Science Campus, delivered an insightful talk on the diverse projects undertaken by the campus. He discussed past and ongoing work applying data science to social policy and international development, including ship traffic analysis in key maritime passages, privacy-preserving record linkage, and transformation of Ghana’s statistical service. The seminar highlighted the campus’s role in using data science to tackle global challenges and develop innovative solutions for societal issues.

                  • Codie Wood, Compass CDT cohort 4

Speaker: Dr Timothy Green, Data Scientist, Data Science Campus, Office for National Statistics

www.ons.gov.uk | datasciencecampus.ons.gov.uk 

The Office for National Statistics is the UK’s largest independent producer of official statistics and the recognised national statistical institute of the UK. Our main responsibilities are collecting, analysing and disseminating statistics about the UK’s economy, society and population.
At the Data Science Campus, we apply data science, and build skills, for public good across the UK and internationally. Our focus is to investigate the use of new data sources and data science tools, methods and practices to create new understanding and improve decision-making for public good.
In this seminar I will give an overview of the broad variety of work done across the Data Science Campus and discuss some of my past and active work in applying data science in social policy and international development.
Dr Timothy Green has a background in physics, with a PhD in Observational Astrophysics, where he studied the properties of the giant galaxies in the centre of clusters of galaxies. He has since worked at the British Council, introducing the teaching of coding into public libraries in Bangladesh, before joining the ONS Data Science Campus, as a data scientist, in 2021.

Schedule:

  • 2.00 – 3.00pm Seminar + Q&A (Rm. G.07, Fry Building)
  • 3.00 – 4.00pm Networking session (Common Room, Fry Building)

 

Compass Guest Lecture: Dr Vincenzo Gioia and Professor Ruggero Bellio

We are delighted to announce the upcoming Compass Guest Lecture with Dr Vincenzo Gioia (University of Trieste) and Professor Ruggero Bellio (University of Udine).

Schedule

11am – 12pm: Scalable Estimation of Probit Models with Crossed Random Effects, Professor Ruggero Bellio

1 – 2pm: Additive Covariance Matrix Models: Modelling Regional Electricity Net-Demand in Great Britain, Dr Vincenzo Gioia

Abstracts

Scalable Estimation of Probit Models with Crossed Random Effects
Professor Ruggero Bellio, Department of Economics and Statistics, University of Udine (Italy)

This talk illustrates a scalable approach to mixed effects modeling with a probit link and a crossed random effects error structure. Random effects with a crossed structure arise often in social and business applications, a notable setting being that of electronic commerce, with random effects related to customers and purchased items, respectively. In sparsely sampled crossed data the computation for both frequentist and Bayesian estimation can easily grow superlinearly with respect to the sample size, which severely limits the use of these models for very large settings. The proposed method belongs to the class of composite likelihood estimators, and entails the fit of three misspecified reduced models. The resulting estimator is consistent and has an overall computational cost linear in the number of observations. This is a joint work with Art Owen and Swarnadip Ghosh, Stanford University, and Cristiano Varin, Ca’Foscari University of Venice.

 

Additive Covariance Matrix Models: Modelling Regional Electricity Net-Demand in Great Britain
Dr Vincenzo Gioia, Department of Economics, Business, Mathematics and Statistics University of Trieste (Italy)

Forecasts of regional electricity net-demand, consumption minus embedded generation, are an essential input for reliable and economic power system operation, and energy trading. While such forecasts are typically performed region by region, operations such as managing power flows require spatially coherent joint forecasts, which account for cross-regional dependencies. Here we forecast the joint distribution of net demand across the 14 regions constituting Great Britain’s electricity network. Joint modelling is complicated by the fact that the net-demand variability within each region, and the dependencies between regions, vary with temporal, socio-economical and weather-related factors. We accommodate for these characteristics by proposing a multivariate Gaussian model based on a modified Cholesky parametrisation, which allows us to model each unconstrained parameter via an additive model. Given that the number of model parameters and covariates is large, we adopt a semi-automated approach to model selection, based on gradient boosting. In addition to demonstrating that adopting a covariate-dependent covariance matrix model leads to substantial forecasting performance improvements, comparable to those obtained by using a full rather than a diagonal static covariance matrix, we explore the model output via accumulated local effects and other visual tools to get insights into how the covariates affect net-demand variability and dependencies. This is a joint work with Matteo Fasiolo, University of Bristol, Jethro Browell, University of Glasgow, and Ruggero Bellio, University of Udine.

The Smith Institute: DataScience@work seminar

When: Friday 7 July, 2-3pm with networking in the Common room thereafter

Location: Room 2.04, Fry Building

Abstract:

The Digital Revolution, Radio Waves and Data Science

Over the past three decades, a whole new sector has grown up around the creation and delivery of mobile data services, which is still growing.  A major challenge has been to provide the capacity in the electromagnetic spectrum to keep pace with demand, and this task has fallen largely on the desks on government regulators.  In addressing it, they have drawn upon the experience of physicists, economists, computer scientists and mathematicians, leading to at least one Nobel Prize along the way and many underpinning scientific advances.  The driver for this progress has been the downstream economic benefits, which are hard to estimate accurately, but amount in financial terms to multiple tens of billions annually.  In this talk, I will survey some of the scientific highlights that have enabled this ‘digital revolution’, concentrating on those that have a strong intersection with data science.

Dr Robert Leese is Chief Technical Officer at the Smith Institute and also Fellow in Mathematics at St Catherine’s College, Oxford.  Beginning in the UK, about 15 years ago, he has worked on many of the auction mechanisms that different governments around the world have used to make spectrum available for new services.

About the Smith Institute

Mathematical approaches coupled with data exploration are key to tackling real-world challenges, to finding the solutions that transform industries and enable societies, businesses and governments to thrive.

From improving the performance of railways to meeting Carbon Zero targetsforecasting crop growth to verifying radio spectrum auctions, the Smith Institute has been tackling their clients’ most critical and complex problems with bespoke solutions for over twenty years.

The possibilities of harnessing mathematical tools are infinite; they enable transformation for their clients across a variety of sectors, applying specialist domain knowledge paired with fresh thinking in transportenergydefencesecurityFMCG and radio spectrum.

For further information about The Smith Institute, see their website.

Compass Conference

We are excited to announce that we will be holding our first Compass Conference on Tuesday 13th September 2022, which will be hosted in the newly refurbished Fry Building, home to the School of Mathematics.

Fry Building
The Fry Building and Voronoi installation

The conference will be a celebratory showcase of the achievements of our students, supervisory teams, and collaborations with industrial partners.

Programme

Exact timings are to be confirmed. Welcome refreshments will be served from 9am with talks to start by 10am. The scheduled talks will finish by 5pm and dinner will finish by 8.30pm.

  • Registration, refreshments and welcome talk
  • Lightening talks: 3 min presentations from Compass PhD students
  • Poster viewing session and networking followed by lunch
  • Research talks:
    • Ed DavisUniversal Dynamic Network Embedding (download slides)
    • Ettore FincatoSpectral analysis of the Gibbs sampler with the concept of conductance
    • Alexander ModellSpectral embedding and the latent geometry of multipartite networks (download slides)
    • Hannah SansfordImplications of sparsity and high triangle density for graph representation learning
    • Michael WhitehouseConsistent and fast inference in compartmental models of epidemics via Poisson Approximate Likelihoods
    • Alessio ZakariaYour Favourite Optimizer may not Converge: Click here to see more
  • Special guest lecture: John Burn-Murdoch, Interactive Data Journalist at the Financial Times.

 

For further information, please contact compass-cdt@bristol.ac.uk

Skip to toolbar