**Prof. Larry A. Wasserman ****(Carnegie Mellon University)**

**Title:**** Causal inference for epidemic models**

**Abs**: The effect of an intervention on an epidemic is usually estimated by adding a variable representing the intervention to an epidemic model. This approach has two issues: the causal effect is not easily extracted and, furthermore, the causal effect is not variation dependent from other parameters in the model. I will discuss some approaches to deal with this in the context of our recent work on estimating the effect of social mobility on deaths from Covid-19. We propose a semiparametric marginal structural model motivated by an epidemic model. We estimate the counterfactual time series of deaths under interventions on mobility. We conduct several types of sensitivity analyses. We find that the data support the idea that reduced mobility causes reduced deaths, but the conclusion comes with caveats. There is evidence of sensitivity to model misspecification and unmeasured confounding which implies that the size of the causal effect needs to be interpreted with caution. While there is little doubt that the effect is real, our work highlights the challenges in drawing causal inferences from pandemic data. This is joint work with Matteo Bonvini, Edward Kennedy and Valerie Ventura.

**Bio**: Larry A. Wasserman is a professor in the Department of Statistics & Data Science and the Machine Learning Department at Carnegie Mellon University. Wasserman received his Ph.D. from the University of Toronto in 1988. He received the COPSS Presidents' Award in 1999 and the CRM-SSC Prize in 2002. He was elected a fellow of the American Statistical Association in 1996, of the Institute of Mathematical Statistics in 2004, and of the American Association for the Advancement of Science in 2011. He was elected to National Academy of Sciences in May, 2016.

*Research Areas: Bayesian methodologies, High dimensional regression, Manifold learning, Causal inference*

**Plenary Talk 2**

**Prof. Arnaud Doucet ****(Oxford University & DeepMind****)**

**Title:**** From denoising diffusion models to diffusion Schrodinger bridges - applications to generative modeling and inference **

**Abs**: In this talk, we will first review denoising diffusion models, a powerful class of generative models at the core of the powerful text-to-image systems Imagen and Dall-E-2. These models provide state-of-the-art results, not only for unconditional simulation, but also when used to sample from complex posterior distributions arising in a wide range of inverse problems such as image inpainting or deblurring. A limitation of these models is that they are computationally intensive at generation time as obtaining each sample requires simulating a non-homogeneous diffusion process over a long time horizon. We will then show how a a Schrodinger bridge formulation leads to theoretically grounded algorithms shortening generation time which are complementary to other proposed acceleration techniques. We demonstrate this novel methodology on various applications including image super-resolution, density estimation on manifolds and optimal filtering for state-space models.

**Bio**: Arnaud Doucet obtained his PhD degree in 1997 from University Paris-Sud Orsay. Ever since he has held faculty positions at Melbourne University, Cambridge University, the University of British Columbia and the Institute of Statistical Mathematics. He joined the department of Statistics of Oxford University in 2011 where he is currently Professor. Since 2019, he is also a Senior Research Scientist at Google DeepMind. He was an Institute of Mathematical Statistics (IMS) Medallion Lecturer in 2016, was elected an IMS Fellow in 2017 and was awarded the Guy Medal in Silver from the Royal Statistical Society in 2020.

**ÃÖÀçÈ£ (µå¶ó¸¶¾ØÄÄÆÛ´Ï Founder & CEO)**

**Title:**** ¸íÇÔ¾Û ¸®¸â¹ö°¡ ±×¸®´Â ¹Ì·¡ **

**Bio**:

- (Çö) ¸®¸â¹ö(µå¶ó¸¶¾ØÄÄÆÛÆ¼ Founder & CEO)

- (Àü) º¸½ºÅÏÄÁ¼³ÆÃ ±×·ì (BCG) ÄÁ¼³ÅÏÆ®

- (Àü) µô·ÎÀÌÆ®ÄÁ¼³ÆÃ ÄÁ¼³ÅÏÆ®

- KAIST ÀüÀÚ°øÇÐ ÇÐ»ç