Daniela de Albuquerque

Daniela de Albuquerque

MD-PhD candidate developing deep generative models for neuroscience and biomedical data.

Publications

Inflationary Flows: Calibrated Bayesian Inference with Diffusion-Based Models

Published in ArXiv (preprint), 2024

In this project we exploited a previously established connection between the stochastic and probability flow ordinary differential equations (pfODEs) underlying Diffusion-Based Models (DBMs) to derive a new class of models, inflationary flows, that uniquely and deterministically map high-dimensional data to a lower-dimensional Gaussian distribution via ODE integration.

Recommended citation: De Albuquerque, D., & Pearson, J.(2024). "Inflationary Flows: Calibrated Bayesian Inference with Diffusion-Based Models." arXiv: 2407.08843 [cs, stat]
Download Paper

Deep Generative Analysis for Task-Based Functional MRI Experiments

Published in Journal of Machine Learning Research, 2021

In this work we follow a Generalized Additive Model (GAM) approach to construct a 3D convolutional VAE capable of more flexibly modeling entire brain volumes, while preserving interpretability.

Recommended citation: De Albuquerque, D., Goffinet, J., Wright, R., & Pearson, J. (2021). "Deep Generative Analysis for Task-Based Functional MRI Experiments." Journal of Machine Learning Research. 149(1-29).
Download Paper

The effects of distraction on brain processes underlying signal detection

Published in Neuropsychologia, 2016

This work leverages the high temporal resolution of EEG to study the neural consequences of a global, continuous distractor on signal-detection processes.

Recommended citation: Demeter, E., DeAlbuquerque, D., & Worldorff, M. (2016). "The Effects of ongoing distraction on the neural processes underlying signal detection." Neuropsychologia. 89(335-343).
Download Paper