2023/10/27 –, track 3
We introduce a novel design approach to design, develop, deploy, and maintain AI/ML tools. This real-world research case study in the healthcare domain, to solve societal issues, can provide actionable methodology for any DevOps task using Python. The presented design methods demonstrate how Python is used in all stages of research with data collected from 60 million patients in United Kingdom.
In healthcare, the use of machine learning (ML) and artificial intelligence (AI) is becoming popular than ever, and Python is in the forefront when it comes to the programming language of choice.
In this talk, we outline the use of a design and implementation framework to aid developers to design Human-Centered AI (HcAI) solutions in healthcare. We provide examples from our research to showcase challenges and best practices of using large healthcare datasets such as CPRD in the UK consisting of data collected from 60 million patients, where we analyse and develop AI/ML models using Python. This will allow the audience to get a good understanding of the challenges involved in solving technical and societal issues such as in healthcare. It will demonstrate as a case study of research in United Kingdom as to how widespread adoption of Python, combined with design innovation can lead to accountable, transparent, and safe use of AI/ML tools.
Participants of the talk will be able to appreciate how this approach can make their design, development, deployment, and maintenance well organized and sustainable. Beginners as well as advanced users are often not familiar with the real-world utility perspectives of AI/ML development, and this talk will provide the international community with actionable take-home ideas that they can use in their projects. It will demonstrate how using Python can make even such data-driven complicated tasks more manageable.
The talk will also be accompanied by a poster, which will provide the audience to interact with the speakers beyond the talk. Furthermore, we open up the Q&A session to include a collaborative platform at an international level for anyone from the audience who might be interested to engage in our research work, for future collaboration.
Talk outline and organization
- Introduction to Human-Centered AI Design using Python
(5 mins) - Examples from the read-world research settings
(2 mins) - Challenges, best practices, and Python tools used to research on database with 60 million patients in United Kingdom
(3 mins) - How to use the Health AI Implementation Framework in research
(5 mins) - Interactive Demo: Examples from ongoing research work using Python
(10 mins) - Q&A from the audience
(5 mins)
Audience
This talk is aimed for Python beginners as it provides design and implementation guidance in AI/ML development, deployment, and maintenance. Advanced users will also benefit from this approach and be able to develop solutions using this implementation framework. The talk would also be suited for policy makers and project + product management team members so that they can use the approach in their companies. Researchers in this field can also take away actionable perspectives for research studies. Although the main example is in healthcare, the design principles described in the talk is applicable to any large scale AI/ML development.
Sami is a healthcare professional and digital health designer currently reading for a doctoral degree, DPhil in Primary Health Care, at the University of Oxford. His research looks at the characteristics and mechanisms of using human centred AI design to address multiple long-term conditions (multimorbidities) at the intersection of the digital transformation in healthcare.
Previously, he worked as a Research Fellow at the Berlin Institute of Health at Charité University Hospital Berlin with the Core Interoperability Unit. Prior to that he worked as a Healthcare Researcher with the WHO and the Medical Faculty of Maastricht University, on a project related to SDG 3 – resulting in the WHO Community Engagement Health Promotion Guide.
With his background in healthcare and digital health, his areas of research interests are in medical AI Ethics, explainable AI, health design innovation, patient data privacy, digital health regulation, and the application of artificial intelligence in healthcare systems.
Sami has researched on 3D printing for personalised medicine and has recently developed the Healthcare Differential Privacy Framework to make privacy-preserving data sharing viable for medical research and machine learning. Sami has also worked at Apple Inc. for 8 years in different roles ranging from sales and product development to training new employees and designing company-internal solutions.
Samara Sharmeen is a resident neurosurgeon working in Berlin, and a PhD candidate at the RWTH Aachen University. Her current research is on prognostic clinical outcome and 3D-CT volumetric planning before conventional complex two-stage haepatectomy versus Associating Liver Partition and Portal vein Ligation for Staged haepatectomy – ALPPS. Samara is interested in use of emerging technologies to aid clinical decision making. She has been collaborating in digital health projects since 2019, attending PyCon and PyConJP, and continues to create awareness of Python to her clinical peers. Furthermore, Samara has worked at Apple Inc. for over 5 years as a Product Specialist.