13:30 – 14:30 – Dr. Michela Iezzi (Bank of Italy)
Title: Privacy-preserving computation in the Bank of Italy: a possible adoption roadmap for Homomorphic Encryption
Abstract: Privacy has gained a growing interest due to the increasing and unmanageable amount of confidential data produced. Financial institutions collect, control and process sensitive data, personally or commercially. Concerns about the possibility of sharing data with third parties to gain fruitful insights beset enterprise environments; value resides in data and the intellectual property of algorithms and models that offer analysis results. This impasse locks the availability of high-performance computing resources in the “as-a-service” paradigm and the exchange of knowledge with the scientific community in a collaborative view. Privacy-enhancing Technologies (PETs) are becoming a practical solution for secure and private information sharing. One such PET, Homomorphic Encryption (HE), allows the computation over encrypted data and thus protects the confidentiality of the information while increasing its utility. Against the background of interesting use cases for the Central Bank of Italy, this seminar focuses on how Homomorphic Encryption can be employed to design and develop privacy-preserving enterprise applications and investigates the current position regarding the use of HE in the financial sector.
15:00 – 16:00 – Dr. Saifuddin Syed (University of Oxford)
Title: Scalable Bayesian Inference with Non-Reversible Parallel Tempering
Abstract: Markov chain Monte Carlo (MCMC) methods are the most widely used tools in Bayesian statistics for making inferences from complex posterior distributions. For challenging problems where the posterior is high-dimensional with well-separated modes, MCMC algorithms can get trapped exploring local regions of high probability. Parallel tempering (PT) tackles this problem by delegating the task of global exploration to a tractable reference distribution (e.g. prior) which communicates to the target (e.g. posterior) through a sequence of parallel MCMC algorithms targeting distributions of increasing complexity to the target.
The classical approach to designing PT algorithms relied on a reversibility assumption, making PT challenging to tune and even deteriorating performance when introducing too many parallel chains. This talk will introduce a new non-reversible paradigm for PT (NRPT) that dominates its reversible counterpart while avoiding the performance collapse endemic to reversible PT methods. We will then establish near-optimal tuning guidelines, an efficient black-box methodology scalable to GPUs. Finally, I will discuss come recent application of NRPT in various scientific domains, including cancer genomics, nuclear fusion, political science, and astronomy.
Refreshments available between 14:30 – 15:00