Simpósio Brasileiro de Telecomunicações e Processamento de Sinais (SBrT) is the most important brazilian symposium on telecommunications and signal processing. In the 2024 edition, that happened from October 1th to 4th, I had 1 paper published with my colleagues from Sidi and 3 papers published with Prof. Magno T. M. Silva and students from University of São Paulo:
Paper “On the Impact of Random Node Sampling on Adaptive Diffusion Networks”
This paper includes some of the results obtained by Daniel G. Tiglea during the period he was working to obtain the Ph.D. Degree.
Conference paper “Can Adaptive Diffusion Networks Do Better with Less Data?”
In this paper, published with Dr. Daniel G. Tiglea, and Prof. Magno Silva, we analyze the performance of an algorithm for adaptive diffusion networks that controls the number of nodes sampled per iteration based on the estimation error. Our model shows that ceasing the sampling of the nodes when the estimation error is sufficiently low can slightly improve the steady-state performance.
The PDF file can be obtained in this page.
Paper “Adaptive diffusion networks: An overview”
In this paper, published with Dr. Daniel G. Tiglea, and Prof. Magno Silva, we go through a comprehensive overview of adaptive diffusion networks, from the first papers published on the subject to state-of-the-art solutions and current challenges.
The PDF file can be obtained in this page.
Papers published in SBrT 2023
Simpósio Brasileiro de Telecomunicações e Processamento de Sinais (SBrT) is the most important brazilian symposium on telecommunications and signal processing. In the 2023 edition, that happened from October 8th to 11th, I had 1 paper published with my colleagues from Sidi and 2 papers published with Prof. Magno T. M. Silva and students from University of São Paulo:
Paper “Reducing the Communication and Computational Cost of Random Fourier Features Kernel LMS in Diffusion Networks” published in ICASSP 2023
In this paper, published with PhD candidate Daniel G. Tiglea, Prof. Luis Azpicueta-Ruiz, and Prof. Magno Silva, we propose an extension of the sampling algorithm presented in “A Low-Cost Algorithm for Adaptive Sampling and Censoring in Diffusion Networks” to the adaptive kernel diffusion networks based on random Fourier features.
The idea is to locally adapt the number of nodes censored according to the estimation error, reducing the computational complexity in nonlinear scenarios.
The PDF file can be obtained in this page.
Linear Algebra in Python: Matrix Inverses and Least Squares
Linear Algebra in Python: Matrix Inverses and Least Squares is the second part of the tutorial on
Read More
scipy.linalg, following Working With Linear Systems in Python With scipy.linalg, published on Real Python.
Papers published in SBrT 2022
Simpósio Brasileiro de Telecomunicações (SBrT) is the most important brazilian symposium on telecommunications and signal processing. In the 2022 edition, that happened from September 25th to 28th, we had 2 papers published:
New paper: A Variable Step Size Adaptive Algorithm with Simple Parameter Selection
We propose a normalized least mean squares algorithm with variable step size. Unlike other solutions, it has low computational cost, only three parameters that are simple to choose, and its steady-state performance can be easily predicted.
New paper: An Adaptive Algorithm for Sampling over Diffusion Networks with Dynamic Parameter Tuning and Change Detection Mechanisms
Recently, we proposed a sampling algorithm for diffusion networks that locally adapts the number of nodes sampled according to the estimation error. In this paper, we extend the results, proposing some improvements to the algorithm.