AKASH SEDAI / RESEARCH AND MODELLING
AS. Play arcade ↗
Machine learning   ★   Quantum-inspired models   ★   Financial modelling   ★   Welcome to my homepage

ML and Statistics

Akash
Sedai. PhD

Associate Lecturer

UCL Institute of Finance & Technology

Exploring interconnected systems through machine learning and quantum-inspired models of uncertainty.

01 / Research

Machine learning across
interconnected systems.

My research focuses on machine learning and quantum-inspired models, with an emphasis on learning from varied data and representing uncertainty.

01

Statistical learning

Probabilistic modelling, and generative methods for high-dimensional data.

Related work
02

Quantum-inspired
models

Exploring quantum-inspired approaches to learning, representation, and modelling systems with interacting parts.

Related work

FROM THE RESEARCH NOTEBOOK

Technical notes

IDEAS / METHODS / APPLICATIONS

Notes on modelling choices, geometric intuition, and the assumptions worth looking twice at.

01 / GAUSSIAN PROCESSESOCTOBER 2026

Compress the function,
or sparsify its dependencies?

Some thoughts on inducing points, conditional dependence, and choosing between SVGP and Vecchia approximations.

02 / ROBUSTNESS & SENSITIVITYOCTOBER 2026

Which observations change a GP's hyperparameters?

Response perturbations, marginal-likelihood curvature, and why a large residual does not necessarily imply large parameter influence.

03 / MULTI-TASK LEARNINGOCTOBER 2026

Correlated tasks do not always improve a multi-task GP

A worked explanation of when cross-task contributions cancel, and what changes with noise, incomplete observations and shared fitting.

04 / QUANTUM-INSPIRED MODELSOCTOBER 2026

Classical data, non-commuting lenses

Observation contexts, order-dependent probabilities, and how ordinary data can enter a quantum-inspired latent model.

02 / Publications

Papers & preprints.

03 / Contact

Let's exchange
ideas.

For academic correspondence and research enquiries.