The Statistical Methods for Health Equity Series is a monthly online series co-hosted by the Data Science for Health Equity community, the Alan Turing Institute Health Equity Interest Group, and the Department of Statistical Science at University College London.

---

For our next webinar, we're delighted to welcome Nyalleng Moorosi from the Distributed AI Research Institute (DAIR) who will be joining us to explore some discussions around understanding how we can build models which centre populations often regarded as peripheral. Specific details on the topic are as below:

Topic: Documenting Health Datasets - Incentives, Transparency, Audits and Inclusion

Abstract:

Applications of machine learning for health have shown great promise. With the onset of deep learning, especially in computer vision, diagnostic medical fields such as radiology and histology have to contend with algorithms that can perform better than practitioners. In Electronic Health Systems, AI has been used to optimize the allocation of health products/services, an issue that became even more important with the onset of COVID-19. And all this is possible because of the large models and the large datasets. However, along with these successes, worrying documentation of failures of these artificial intelligence systems is seen when they are tasked with allocating health services to minority populations or detecting certain illnesses in specific populations. Additionally, it is seen that even in fields seemingly devoid of human influence such as radiology, machine learning systems can infer social categories such as race from samples and therefore influence downstream. In this talk, Nyalleng will present work on why data documentation is crucial, the difficulty of incentivizing data users and producers to create detailed data artifacts, and how we might start to think about the role of diversity in creating ML for Health systems.

Nyalleng is a senior researcher at DAIR since December 2022. Before DAIR she was a research software engineer at Google AI, where she was one of the first employees at the Google Africa research lab. She has also been a senior researcher at the South African Council for Scientific and Industrial research, where she worked closely with government and academic institutions to develop a diversity of products for private and public institutions. Outside of formal work she is involved in efforts to democratize AI; she is a founding member of the Deep Learning Indaba, the largest machine learning consortium of AI/ML practitioners in Africa, a member of A+ Alliance an international coalition that seeks to not only detect, but correct, gender bias in Artificial Intelligence.

--

Please direct any questions about the webinar series to Dr Brieuc Lehmann at b.lehmann@ucl.ac.uk.