NORA Annual Conference 2022 - Keynotes
Uli Sattler
Professor in the Information Management Group within the Department of Computer Science of the University of Manchester
Uli Sattler is a professor at the University of Manchester, working in logic-based knowledge representation, Description Logics, and ontology engineering. Together with colleagues, she designed the family of Description Logics underpinning ontology languages such as OWL and decision procedures for relevant reasoning problems of these logics. She has been working on a range of standard and novel reasoning problems that are important for ontology engineering and usage such as module extraction and decomposition, entailment explanation, and mining axioms from data. Uli completed her PhD under the supervision of Franz Baader at RWTH Aachen and her habilitation at TU Dresden.
Title of the keynote:
Background Knowledge, graphs, and ontologies: potential and challenges for using knowledge in machine learning
In recent years, the power of neural models of AI in machine learning, computer vision, and natural language processing has massively increased due to increased computing power and availability of data. There are, however, still significant challenges in the area, in particular related to issues of transparency, trust, and explainability as well as the reliance on large amounts of training data and computation power. While all these are also addressed by advanced ML techniques, in this talk I want to highlight the potential and challenges of semantic technologies to address these issues.
Knowledge bases, ontologies, knowledge graphs, and related (reasoning and other) tools are semantic technologies that have been developed over the last decades in symbolic AI. They are means to capture and work with rich domain knowledge (e.g., in medicine) in a logically well-founded and computer processable ways, and are being widely used in knowledge-rich areas, e.g., to maintain and interact with medical or bio-chemistry ontologies.
In this talk, I want to discuss how these ontologies can be used to address some of the above issues in ML. For example, we explore different ways to equip neural models with relevant background knowledge from ontologies as prior knowledge, so that models can perform better with less training data, and so that we can inspect embeddings through the lens of our background knowledge. I will also talk about using ontologies to evaluate performance of classifiers in a way that takes into account the seriousness of errors, thereby going beyond simple ‘correct/incorrect’ analysis and providing deeper insight into the performance of our models.
Evandro Fei Fang
Associate Professor - Department of Clinical Molecular Biology, University of Oslo
The Evandro F. Fang group are investigating the molecular mechanisms of human aging, age-predisposed Alzheimer’s disease (AD), and artificial intelligence (AI)- based drug development for pathological ageing and AD. After finished his Ph.D. training in Biochemistry with Chair Professor Tzi Bun Ng at The Chinese University of Hong Kong in 2012, Fang started a 5-year postdoctoral fellowship at the National Institute on Aging USA with Professor Vilhelm Bohr on DNA repair and mitophagy in accelerated ageing, and with Professor Mark Mattson on NAD+ in neurodegeneration. In September 2017, he established his independent laboratory at The University of Oslo, Norway (https://evandrofanglab.com/). His laboratory is focusing on the molecular mechanisms of how cells clear their damaged and aged mitochondria, a process termed “mitophagy”, as well as the roles of mitophagy in AD. NAD+ is a fundamental molecule in life and health, and his group is investigating the molecular mechanisms on how NAD+ alleviates ageing and agepredisposed neurodegeneration, especially AD. He is fascinated with and actively engaged in moving his laboratory findings to translational applications, as involved in 4 NAD+ -based clinical trials, with the overarching goal to establish novel and safe biological approaches to promote longer and healthier human lives.
He has published over 70 papers in international peer-reviewed journals including papers in Cell, Cell Metabolism, Nature Reviews MCB, and Nature Neuroscience. He has received several awards including The NIH Fellows Award for Research Excellence 2014, 2015, the Butler-Williams Scholar on Aging 2016 (USA), a FRIMEDBIO Young Research Talent 2017(Norway), a finalist of the 2017 ERC Starting grant, and the Norwegian Cancer Society Pink Ribbon 2020 Award. On 28th Feb 2020, he received the DKNVS 'scientific award to young scientist in the natural sciences for 2020’ (DKNVS’ vitenskapelige pris til yngre forskere innen naturvitenskap for 2020), by The Royal Norwegian Society of Sciences and Letters / Det Kongelige Norske Videnskabers Selskab (DKNVS) (https://evandrofanglab.com/2020/01/09/no-age-members-honor-dknvs-awards/).
Title of the keynote:
The NAD+-mitophagy axis in ageing and Alzheimer’s disease and AI-based drug development
There were 962 million elderly (60+) people globally in 2017, and this number will rise to 2.1 billion in 2050, bringing formidable healthcare and socio-economic challenges; similarly, China is also entering into an ageing society1, 2. Ageing is arguably the highest risk factor for numerous human diseases, thus understanding the molecular mechanisms of human aging holds the promise of developing interventional and therapeutic strategies for many diseases simultaneously, promoting healthy longevity. Accumulation of damaged mitochondria is a hallmark of aging and age-related neurodegeneration, including Alzheimer’s disease (AD). However, the molecular mechanisms of impaired mitochondrial homeostasis and their relationship to AD are still elusive. Mitophagy is the cellular self-clearing process of damaged and superfluous mitochondria, and therefore plays a fundamental role in maintaining neuronal function and survival3. We hypothesize that age-susceptible defective mitophagy causes accumulation of damaged mitochondria, which in combination with the two AD-defining pathologies, Aβ plaques and tau tangles, further exacerbates AD progression. Restoration of mitophagy through upregulation of cellular NAD+, a primary molecule in human health and life, and genetic approaches, forestalls pathology and cognitive decline in C. elegans and three mouse models of AD and improves mitochondrial function in the AD iPSC neurons4, 5. We are now involved in more than 5 clinical trials on the use of NAD+ precursors to treat AD, and premature ageing diseases, among others6. Additionally, we are using artificial intelligence (AI) to propel drug screening and drug design targeting AD and ageing pathways7.
References
1. Fang, E.F. et al. A research agenda for ageing in China in the 21st century (2nd edition): Focusing on basic and translational research, long-term care, policy and social networks. Ageing Res Rev 64, 101174 (2020).
2. Fang, E.F. et al. A research agenda for aging in China in the 21st century. Ageing Res Rev 24, 197-205 (2015).
3. Aman, Y. et al. Autophagy in healthy ageing and disease. Nat Aging 1, 634-650 (2021).
4. Fang, E.F. et al. Mitophagy inhibits amyloid-beta and tau pathology and reverses cognitive deficits in models of Alzheimer's disease. Nat Neurosci 22, 401-412 (2019).
5. Lautrup, S., Sinclair, D.A., Mattson, M.P. & Fang, E.F. NAD(+) in Brain Aging and Neurodegenerative Disorders. Cell Metab 30, 630-655 (2019).
6. Reiten, O.K., Wilvang, M.A., Mitchell, S.J., Hu, Z. & Fang, E.F. Preclinical and clinical evidence of NAD(+) precursors in health, disease, and ageing. Mech Ageing Dev, 111567 (2021).
7. Xie, C. et al. Amelioration of Alzheimer’s disease pathology by mitophagy inducers identified via machine learning and a cross-species workflow. Nat Biomed Eng (In press).
Ulrich Paquet
Research Scientist at DeepMind
Ulrich Paquet is a senior staff research scientist at DeepMind, one of the world's foremost AI companies. His most recent work on AlphaZero include a collaboration with Vladimir Kramnik on how small rule adjustments to chess would alter the game at a superhuman level. He is deeply interested in how AlphaZero discovers and acquires chess knowledge automatically, and has done research showing how it represents human chess concepts.
He holds a PhD in machine learning from the University of Cambridge. Ulrich joined DeepMind in 2016, and has worked in tech companies such as Microsoft, Imense, VocalIQ and Apple before. He was on the Xbox One launch team, where he co-developed the core of the recommender system. Having grown up in South Africa, he cares deeply about the growth of mathematical sciences in Africa. In his spare time, he is a co-founder and co-director of Deep Learning Indaba, a grassroots initiative to strengthen machine learning in Africa. To date, it has run AI schools in almost 30 countries in Africa, with many flowers now blooming between the grass roots!
Title of the keynote:
AlphaZero: A Retrospective and Future Story