Publications
Scientific research behind our methodology
Our work is grounded in mathematical, statistical and computational research. Through academic and industry collaborations, we develop and validate methods for complex biomedical data problems. Our publications demonstrate the scientific foundation behind our approach — from high-dimensional statistics and overfitting to survival analysis, patient risk prediction and federated inference.
From mathematical theory to biomedical applications
Rather than presenting all publications equally, we highlight research that directly connects to our current solutions.
Publications by Research Area
High-dimensional statistics and overfitting correction
-
Practical debiasing with the Covariant Prior in the proportional regime when p < n, 2025
-
Observable asymptotics of regularized Cox regression models with standard Gaussian designs, 2025
-
Replica analysis of overfitting in regression models for time-to-event data: The impact of censoring, 2024
-
Exact results on high-dimensional linear regression via statistical physics, 2021
-
Replica analysis of overfitting in generalized linear regression models, 2020
-
Replica analysis of overfitting in regression models for time-to-event data, 2017
Latent Heterogeneity and Competing Risks
Statistical and Computational Methodology
-
Delayed kernels for longitudinal survival analysis and dynamic prediction, 2024
-
Fast and accurate recurrent event analysis for genome-wide association studies, 2023
-
Retarded kernels for longitudinal survival analysis and dynamic prediction, 2021
-
Accurate Bayesian data classification without hyperparameter cross-validation, 2019
-
Bayesian clinical classification from high-dimensional data: Signatures versus variability, 2016
Related research
Research Behind Our Solutions
Connecting publications to applications
Our scientific research directly informs the development of our software solutions.
spsSIGNATURE
-
High-dimensional statistics & overfitting
spsMOSAICS
-
Discovery of latent patient subgroups and potential responders
Bayesian Federated Inference
-
Statistical analysis across distributed datasets without pooling patient-level data
Research through collaboration
We collaborate with academic researchers, healthcare organisations and industry partners to develop and apply new statistical methodology to challenging biomedical questions.These collaborations allow us to connect:
-
Fundamental research
-
Method development
-
Validation
-
Biomedical application