The Rise of AI in the Healthcare Ecosystem

home | speakers | programme | registration


Dr Firat Tekiner (Senior Product Manager, Google)

What are the things that AI can help with in the healthcare ecosystem?

AI has the potential to change how we do things and is likely to have an impact on our productivity. AI is becoming more accessible to different professionals, whether you have ML experience or not. Healthcare is no different to this and it will help organisations to support health professionals with dedicated tools.

Firat Tekiner PhD Reinforcement Learning, MBA, is a Senior Staff Product Manager in Data Analytics and AI at Google Cloud. Firat is a leader with nearly 20 years of experience in developing new products, designing and delivering bespoke information systems for some of the world’s largest research, education, telecommunications, finance and retail organisations. Following roles within National Supercomputing Services and National Centre for Text Mining, he worked as a senior consultant at Ab Initio software prior to joining Google. Firat is the co-author of "Architecting Data and Machine Learning Platforms", Oct2023, ISBN:1098151615. He has over 50 publications in the areas of Parallel Computing, Big Data, Artificial Intelligence and Computer Communications.

Dr Nava Maroto (Department of Linguistics Applied to Science and Technology, Universidad Politécnica de Madrid, Spain)

Semantic Deep Learning for One Health: combining Knowledge Graphs and Large Language Models

"One Health" considers public and animal health and the ecosystems interlinking them. PubMed is a repository of 35 million papers (titles and available abstracts), including scientific findings relevant for both public and animal health. How to transform textual information into knowledge graphs (computer-interpretable format) sharing common terminologies (i.e. terms/concepts) is a current challenge. This talk will present a Semantic Deep Learning approach (SemDeep for short) combining Semantic Web (including Knowledge Graphs) and Deep Learning research (including large language models like BERT and ChatGPT).

Dr Nava Maroto is a Lecturer and Researcher at the Department of Linguistics Applied to Science and Technology at the Universidad Politécnica de Madrid (Spain). Her research interests include terminology and neology, both from a theoretical and practical perspective. Over the last years she has collaborated with Dr. Mercedes Argüello-Casteleiro in the development of deep learning algorithms for information retrieval in the field of One Health. She is also Vice-Dean for External Promotion at the School of Telecommunications Engineering at the Universidad Politécnica de Madrid, where she works within a multidisciplinary team. She strives to bridge the gap between engineering and linguistics, and she firmly believes in the power of interdisciplinary cooperation to meet all kinds of challenges.

Dr Julie Klein, Institute of metabolic and cardiovascular disease, Inserm/Université Paul Sabatier, France.

Biomarkers and Personalised Medicine: how can AI be used for Omics Data Integration, Interpretation, and Its Application?

The "democratisation of eScience" is about widening the audience to researchers (e.g. biologists) and practitioners (e.g. clinicians), who are not experts in AI/data science, by providing automatisation of data-oriented/scientific workflows while keeping the human in the loop. For example, the UniProt Knowledgebase (UniProtKB) or the gene ontology are examples of widely used information resources. However, moving from generic e-Infrastructure to AI web/mobile apps usable by domain experts is a challenge. The UniProtKB comprises UniProtKB/Swiss-Prot (more than 500K entries manually curated) and UniProtKB/TrEMBL (more than 200M entries computationally generated records). The UniProtKB in combination with other information sources like MEROPS (the Peptidase Database) can aid biologists investigating biomarkers for long-term conditions, such as Chronic Kidney Disease (CKD). Recent research suggests that 1 in 10 people may have CKD. People with any stage of CKD have a higher risk of developing heart disease.

Dr Julie Klein received her PhD from the University Paul Sabatier, Toulouse, France, in 2009. From 2009 to 2014 she worked as a post-doctoral fellow in the field of proteomics and bioinformatics in chronic kidney disease and was recruited as a researcher at the Renal Fibrosis Lab, INSERM, Toulouse in 2014.

Prof Chris Freeman (Electronics & Computer Science, University of Southampton, UK)

AI in Next Generation Stroke Rehabilitation

As progress in smart textiles, soft robotics and pervasive sensing networks gathers pace, the prospect of truly wearable technology for stroke rehabilitation is fast emerging. However the ability of control systems and machine learning to support this vision is lagging far behind, limiting its potential to restore independence and dignity to people with long term conditions. An overview of current research into advanced control and machine learning strategies for electrical stimulation and robotics will be presented, showing their effectiveness for upper limb stroke rehabilitation. Then a discussion is given of open control, identification, sensing and motor learning problems that must be solved to maximise its effectiveness and allow it to reach the end user.

Prof Chris Freeman is a Professor in Electronics and Computer Science at the University of Southampton, UK. His research interests include the development, application and assessment of iterative learning, repetitive and adaptive controllers for both biomedical and industrial systems. Over the last fifteen years he has developed new healthcare technologies combining robotics and electrical stimulation to enable people with upper limb impairments to perform functional tasks. Over this time he has worked closely with clinicians, patients and carers. These include five clinical trials using technology he has helped develop, as well as numerous smaller studies and user-led design sessions. His focus has been to understand and define clinical problems within an engineering perspective and translate this into usable solutions. His research covers the spectrum from control application, control theory, rehabilitation engineering, biomechanics, clinical studies and user perspectives.

Rik Smithies (HL7 UK)

Machine Readable Healthcare Data Formats using FHIR

A common representation of health data is a crucial foundation for machine learning and AI. HL7 FHIR is a data representation and access specification for all healthcare information. Since its creation 10 years ago it has become the leading global approach to sharing health data, with wide industry support that includes Microsoft, Google, IBM and Apple, as well as most healthcare specific software vendors. An overview of FHIR will be presented, explaining how it works with large scale REST APIs, and integrates with Clinical Systems.

Rik Smithies is an independent HL7 FHIR consultant with over 20 years medical informatics experience. He is an active committer on the FHIR development team, a regular international FHIR speaker and a FHIR Certified Implementer. Rik is currently Technical Chair of HL7 UK - a not-for-profit organization that promotes the use of international healthcare interoperability standards. He was elected as a Fellow of HL7 International in 2022, to mark 15 years of volunteer service. Recent clients include NHS (UK), FDA (US), EMA (EU), Microsoft (US), IBM (US) and Cambridge University (UK).

Dr Jeremy Rogers (Terminology and Classifications Delivery Service, NHS England, UK)

What SNOMED wants to be, how clinicians actually use it and what lurks in large coded clinical datalakes?

Applying AI to healthcare typically requires large corpora of detailed patient-level health information on which to train. In addition to the copious quantities of semi- and unstructured free-text narrative documents written by clinicians to each other, and the many numerical and digital image results generated through investigation, a third and apparently more immediately "machineable" source of rich clinical data has always been the coded portion of the Electronic Patient Record. With increasing volumes of such codings globally now being populated directly by clinicians using SNOMED CT, this talk will review the data capture and analytics problems that SNOMED CT was designed to try and solve, the new problems it has created and what is known about how UK clinicians actually use it. The implications for training AI on repositories of SNOMED CT coded clinical data will be considered.

Dr Jeremy Rogers FFCI MD MB ChB is a Consultant Terminology Specialist within the NHS Information Representation Service, previously within NHS Digital and now part of NHS England's Technology Strategy, Architecture and Standards directorate. Since 2005 he has been developing NHS strategy and solutions to enable SNOMED CT's implementation across Health and Social Care in the UK. He was previously a minimum part-time GP and a Research Associate in Medical Informatics at University of Manchester, where his research interests covered many aspects of clinical knowledge representation and reasoning.

Dr Chris Wroe (Head of Health Informatics, British Medical Journal (BMJ), UK)

Bringing clinical guidelines closer to the bedside with AI techniques

It’s been a long term goal of the healthcare community to close the gap between what is recommended in evidence based guidelines, and what is actually achieved at the bedside. BMJ has a mission to make it easier for clinicians to access these recommendations, but it can be a challenge to seamlessly incorporate into the complex work environment of busy clinicians. Not only that, but how do we support clinicians apply the general recommendations to specific patients?

Building on the increasing availability of data through the use of SNOMED-CT and FHIR, I’ll talk about how related standards such as FHIR Clinical Practice Guidelines and Clinical Quality Language may enable us to take the next step in encoding knowledge as well as data. With the development of large language models, I’ll also cover the promise and challenge of their use in this context.

Dr Chris Wroe is Head of Health Informatics at BMJ. He has been focussing on how to get evidence based recommendations in front of clinicians using informatics standards such as SNOMED-CT and those from HL7. Prior to working at the BMJ, he was a solution specialist at BT Global Services, working on projects deploying integrated clinical systems within London hospitals. He also spent a significant period as a biohealth informatics clinical fellow at Manchester University. His research focused on how ontologies can underpin a range of applications from clinical decision support to genomics research.


Organised by BCS SGAI
The Specialist Group on Artificial Intelligence