Pedro Campos Vieira

Pedro Campos Vieira

I am Pedro, a researcher in Artificial Intelligence and Data Science with a Master's degree in the field. My broad area of interest is Network Science and Machine Learning. Currently I focus on topics at the intersection of geometry, topology, and deep learning. I'm aiming to continue this path through a Ph.D. in Topological and Geometric Deep Learning.

Sitemap

A list of all the posts and pages found on the site. For you robots out there, there is an XML version available for digesting as well.

Pages

✦ ✦ ✦
✦ ✦ ✦
✦ ✦ ✦
✦ ✦ ✦

✦ ✦ ✦

✦ ✦ ✦
✦ ✦ ✦
✦ ✦ ✦
✦ ✦ ✦
✦ ✦ ✦
✦ ✦ ✦
✦ ✦ ✦
✦ ✦ ✦
✦ ✦ ✦
✦ ✦ ✦

✦ ✦ ✦

✦ ✦ ✦

✦ ✦ ✦

✦ ✦ ✦

✦ ✦ ✦

✦ ✦ ✦

✦ ✦ ✦

✦ ✦ ✦
✦ ✦ ✦

✦ ✦ ✦

✦ ✦ ✦

✦ ✦ ✦

Posts

Essentials of Backpropagation in Fully-Connected Feedforward Multilayer Percepetrons

9 minute read

Published:

For the first blog post, I think it is fitting to start with some fundamental topic. Hence, I decided to start with a fundamental topic for mahcine-learning: backpropagation. This is a fundametal mechanism that is widely used in many modern and state-of-the-art methods and deep-learning architectures. So, it is kind of important :smile:. This short view into backpropagation will go over how the mechanism works in simple Fully-Connected Feedforward Multilayer Percepetrons. Even though it remains in how it is appliyed, probably, in a future blog post, I will tackle more complex arquitectures.

✦ ✦ ✦

publications

S+t-SNE - Bringing Dimensionality Reduction to Data Streams

Published in 22nd International Symposium on Intelligent Data Analysis, 2024

We adapt t-SNE to online scenarios!

Recommended citation: C. Vieira, P., Montrezol, J. P., T. Vieira, J., & Gama, J. (2024). S+t-SNE - Bringing Dimensionality Reduction to Data Streams. In I. Miliou, N. Piatkowski, & P. Papapetrou (Eds.), Advances in Intelligent Data Analysis XXII (pp. 95–106). Cham: Springer Nature Switzerland.
Download Paper | Download Bibtex

✦ ✦ ✦

The Dynamic-Loose octree : a spatial index structure towards time-efficiency

Published in Journal peer-reviewed to be published in mid-to-late Winter 2025., 2024

We join the dynamic and the loose enhancement of the octree. Contrarly to previous approaches, we maintain the benefits from both enhancements.

Recommended citation: Carneiro, E., C. Vieira, P., Carvalho, A. V., Amaro, M., The Dynamic-Loose octree : a spatial index structure towards time-efficiency.
Download Paper

✦ ✦ ✦

Studying and Improving Graph Neural Network-based Motif Estimation

Published in arXiv, 2025

We reaproximate motif estimation with GNNs to the traditional motif concept. We also propose reframing the target task to a direct multi-target significance-profile estimation.

Recommended citation: Vieira, P. C., Silva, M. E. P., & Ribeiro, P. M. P. (2025). Studying and Improving Graph Neural Network-based Motif Estimation. doi:10.48550/ARXIV.2506.15709
Download Paper | Download Bibtex

✦ ✦ ✦

talks

✦ ✦ ✦
✦ ✦ ✦