“Some of the most effective learning algorithms are those that combine perspectives from many different models or parameters. This has always seemed a fitting metaphor for effective research. And now ELLIS will provide a new architecture to keep our real-life committee machine functioning --- reinforcing, deepening and enlarging the channels that connect us to colleagues throughout Europe At UCL we're excited to be a part of this movement to grow together. We look forward to sharing new collaborations, workshops, exchanges, joint studentships and more, and to the insight and breakthroughs that will undoubtedly follow. ”
Prof Maneesh Sahani
Director, Gatsby Computational Neuroscience Unit
“Advances in AI that benefit people and planet require global cooperation across disciplines and sectors. The ELLIS network is a vital part of that effort and UCL is proud to be a contributor. ”
Prof Geraint Rees
UCL Pro-Vice-Provost (AI)
In this talk, I provide my perspective on the machine learning community's efforts to develop inference procedures with Bayesian characteristics that go beyond Bayes' Rule as an epistemological principle. I will explain why these efforts are needed, as well as the forms which they take. Focusing on some of my own contributions to the field, I will trace out some of the community's most important milestones, as well as the challenges that lie ahead. Throughout, I will provide success stories of the field, and emphasise the new opportunities that open themselves up to us once we dare to go beyond orthodox Bayesian procedures.
Causal inference is an important tool for answering many scientific questions in medicine. However, many statistical and machine learning methods can only provide association answers. Research in medicine usually involves complex sceneries, and causal inference in complex sceneries becomes more complicated, particularly regarding to causal estimands and their identifiability. In this talk, I discuss how to make causal inference methods in medicine. First, I discuss causal inference methods for randomized trials with concurrent events. Randomized trials are often considered the gold standard design for evaluating causality for new interventions, for example, interventions aimed at improving service delivery, organization, quality, financing, and health care outcomes. In practice, however, many randomized protocol violations can occur for various reasons, and trials that do not fully adhere to protocol, concomitant problems such as truncation by death and rescue medication can occur. In such destructive randomized trials, standard estimation methods are no longer fully valid. In this talk, I discuss the new developments in causal inference in three violations of the ideal randomized scheme. Second, I discuss causal inference in precision medicine. Patients need to choose the best treatment plan according to their own conditions, and doctors need to choose the most suitable patient for personalized treatment according to the characteristics of each treatment plan. The selection of individualized treatment plans in precision medicine involves the intersection of multiple disciplines, such as medicine, probability and statistics, computational mathematics, and applied mathematics. Third, I discuss causal inference methods when the outcome of interest is latent and has to be estimated from data. Finally,I discuss causal inference for recommender systems.