Posts by Collection

Regularised B-splines projected Gaussian Process priors to estimate time-trends of age-specific COVID-19 deaths related to vaccine roll-out

Appeared in Bayesian Analysis, 2021

We fit beta splines to COVID-19 mortality data for each US state and use the fitted curves to estimate future casualties.

Recommended citation: Monod, M., Blenkinsop, A., Brizzi, A., Chen, Y., A.C.C. Perello, C., Jogarah, V., Wang, Y., Flaxman, S., Bhatt, S., & Ratmann, O. Regularised B-splines projected Gaussian Process priors to estimate time-trends of age-specific COVID-19 deaths related to vaccine roll-out, (2021). https://projecteuclid.org/journals/bayesian-analysis/volume-18/issue-3/Regularised-B-splines-Projected-Gaussian-Process-Priors-to-Estimate-Time/10.1214/22-BA1334.full

Adaptively Optimised Adaptive Importance Samplers

Preprint (submitted for publication), 2023

We introduce a new adaptive importance sampling algorithm that uses optimisation to adapt the proposal distribution to obtain a tighter envelope of the target distribution.

Recommended citation: C. A.C.C. Perello, Akyildiz, Ö. D. Adaptively Optimised Adaptive Importance Samplers, (2023). https://arxiv.org/abs/2307.09341

Graph-based mutually exciting point processes for modelling event times in docked bike-sharing systems

Appeared in Stat, 2023

We fit a spatio-temporal Hawkes processes to the London Santander Cycle bike-sharing system and study the resulting model.

Recommended citation: F. Sanna Passino, Y. Che, C. A.C.C. Perello. Graph-based mutually exciting point processes for modelling event times in docked bike-sharing systems, (2023). https://onlinelibrary.wiley.com/doi/full/10.1002/sta4.660

Non-existence of the 2D quadratic Poisson optimal matching

MSc Thesis, 2024

I analyse a recent proof of the fact that a stationary, non-ergodic, quadratic and locally-optimal 2D Poisson matching does not exist.

Recommended citation: C. A.C.C. Perello. Non-existence of the 2D quadratic Poisson optimal matching, (2024).

Optimal Transport with Huber Loss: Barycentres and Robustness

In preparation, 2026

We develop a robust alternative to Wasserstein barycentres using optimal transport with Huber loss, establishing stability, existence, and robustness properties.

Recommended citation: C. A.C.C. Perello, A. González-Sanz. Optimal Transport with Huber Loss: Barycentres and Robustness, (2026+). In preparation.

Sharp Asymptotics for Regularised Optimal Transport

In preparation, 2026

We study sharp small-regularisation asymptotics for a unified family of regularised optimal transport problems, including entropic and $L^p$-regularised optimal transport.

Recommended citation: C. A.C.C. Perello, A. González-Sanz, M. Nutz. Sharp Asymptotics for Regularised Optimal Transport, (2026+). In preparation.

Peer Tutoring

Peer Tutoring, Imperial College London, Department of Mathematics, 2022

I was a peer tutor for the first year mathematics course at Imperial College London. I was responsible for leading weekly tutorials for a group of 5-6 students, helping them develop their problem solving skills and understanding of the course material. I also prepared challenging problems for the students to work on in their own time.

UTA for MATH50003 Numerical Analysis

Undergraduate Teaching Assistant, Imperial College London, Department of Mathematics, 2023

I was an Undergraduate Teaching Assistant (UTA) for MATH50003 Numerical Analysis at Imperial College London, a 2nd year undergraduate course led by Prof. Sheehan Olver. I attended the weekly problems classes and helped students with problems, especially with the Julia programming exercises. I also marked the mock exam for said course. More information can be found on the course’s GitHub page.

Guest lectures

Lecturing, Brown University, 2025

I gave four guest lectures for APMA1650: Statistical Inference I at Brown University.

TA for STAT5264 Stochastic Processes with Applications

Teaching Assistant, Columbia University, Department of Statistics, 2025

I was teaching assistant for STAT5264 Stochastic Processes with Applications, a master’s level course which is mandatory for students in the Mathematics of Finance program at Columbia Univesity. The instructor was Prof. Graeme Baker. Roughly speaking, the course introduces stochastic calculus, Brownian motion and classical mathematical finance, such as delta-hedging and risk-neutral pricing via Girsanov’s formula. A syllabus be found here.

TA for STAT5265 Stochastic Methods in Finance

Teaching Assistant, Columbia University, Department of Statistics, 2026

I was teaching assistant for STAT5265 Stochastic Methods in Finance, a master’s level course which is mandatory for students in the Mathematics of Finance program at Columbia Univesity. The instructor was Prof. Graeme Baker. The course introduces students to advanced applied probability methods in mathematical finance, such as option pricing, optimal control and Feynman-Kac formulae. I also gave two guest lectures for this course. A syllabus be found here.