Dear all,
unfortunately, the next OWABI seminar, initially scheduled on February 27, has been cancelled due to unforeseen circumstances. Ayush Bharti's talk will be rescheduled on the 24th April.
Our next OWABI talk will be given by Meïli Baragattihttps://www.vinifera-euromaster.eu/team/meili-baragatti/ (Assistant Professor Institut Agro Montpellier) on Approximate Bayesian Computation with Deep Learning and Conformal prediction on Thursday the 27th March 2025. More info will follow.
Best, Massimiliano on the behalf of the OWABI Organisers
------ Dr. Massimiliano Tamborrino Reader and WIHEA Fellow Department of Statistics University of Warwick https://warwick.ac.uk/tamborrino ________________________________ From: Tamborrino, Massimiliano Sent: 13 February 2025 12:37 To: abc_world_seminar@listserv.csv.warwick.ac.uk abc_world_seminar@listserv.csv.warwick.ac.uk Subject: OWABI - Ayush Bharti - Feb 27
Dear all,
the next OWABI seminarhttp://www.warwick.ac.uk/oneworldabc is scheduled on Thursday the 27th February at 11am.
I am pleased to inform you that our next speaker is Ayush Bharti (Aalto University), who will talk about "Cost-aware simulation-based inference ", with an abstract reported below.
The talk will be streamed on MS Teams on the OWABI Ms Teams channel OWABI Seminar: One World Approximate Bayesian Inference Seminar | General | Microsoft Teamshttps://teams.microsoft.com/l/team/19%3AdhZ_4e_XLNJzCXPAMzTvT6BZ5KShEETkd_wtTY52VI81%40thread.tacv2/conversations?groupId=9c061d11-f88c-4cee-938f-bf40e7393879&tenantId=09bacfbd-47ef-4465-9265-3546f2eaf6bc
You could join the meeting with the link and details below Join the meeting nowhttps://teams.microsoft.com/l/meetup-join/19%3adhZ_4e_XLNJzCXPAMzTvT6BZ5KShEETkd_wtTY52VI81%40thread.tacv2/1739444032750?context=%7b%22Tid%22%3a%2209bacfbd-47ef-4465-9265-3546f2eaf6bc%22%2c%22Oid%22%3a%2242c962d4-7e3d-42ae-ad13-21a387aadf72%22%7d Meeting ID: 311 178 293 223 Passcode: zg3E8kw3 Important: The virtual lobby has been removed, so everyone should now be able to join the seminar without any authorisation.
Abstract: Simulation-based inference (SBI) is the preferred framework for estimating parameters of intractable models in science and engineering. A significant challenge in this context is the large computational cost of simulating data from complex models, and the fact that this cost often depends on parameter values. We therefore propose cost-aware SBI methods which can significantly reduce the cost of existing sampling-based SBI methods, such as neural SBI and approximate Bayesian computation. This is achieved through a combination of rejection and self-normalised importance sampling, which significantly reduces the number of expensive simulations needed. Our approach is studied extensively on models from epidemiology to telecommunications engineering, where we obtain significant reductions in the overall cost of inference.
Keywords: simulation-based inference, approximate Bayesian computation, neural posterior estimation, neural likelihood estimation, importance sampling
Best, Massimiliano on the behalf of the OWABI Organisers
------ Dr. Massimiliano Tamborrino Reader and WIHEA Fellow Department of Statistics University of Warwick https://warwick.ac.uk/tamborrino