CORTICAL-INSPIRED PLACEMENT AND ROUTING: MINIMIZING THE MEMORY RESOURCES IN MULTI-CORE NEUROMORPHIC PROCESSORS
Oct 2022 - BioCAS'22
Brain-inspired event-based neuromorphic processing systems have been emerging as a promising technology, in particular for bio-medical circuits and systems. However, both neuromorphic and biological implementations of neural networks have critical energy and memory constraints. To minimize the use of memory resources in multi-core neuromorphic processors, we propose a network design approach that takes inspiration from biological neural networks. We use this approach to design a new routing scheme optimized for small-world networks and, at the same time, to present a hardware-aware placement algorithm that optimizes the allocation of resources for small-world network models. We validate the algorithm with a canonical small-world network and present preliminary results for other networks derived from it.
AN ON-CHIP SPIKING NEURAL NETWORK FOR ESTIMATION OF THE HEAD POSE OF THE ICUB ROBOT
Jun 2020 - Frontiers in Neuroscience
In this work, we present a neuromorphic architecture for head pose estimation and scene representation for the humanoid iCub robot. The spiking neuronal network is fully realized in Intel's neuromorphic research chip, Loihi, and precisely integrates the issued motor commands to estimate the iCub's head pose in a neuronal path-integration process. The neuromorphic vision system of the iCub is used to correct for drift in the pose estimation. Positions of objects in front of the robot are memorized using on-chip synaptic plasticity. We present real-time robotic experiments using 2 degrees of freedom (DoF) of the robot's head and show precise path integration, visual reset, and object position learning on-chip. We discuss the requirements for integrating the robotic system and neuromorphic hardware with current technologies.
APPLICATION ON REINFORCEMENT LEARNING FOR DIAGNOSIS BASED ON MEDICAL IMAGE
Jan 2008 – Reinforcement Learning: Theory and Applications, Book edited by Cornelius Weber, Mark Elshaw and Norbert Michael Mayer ISBN 978-3-902613-14-1, pp.424, January 2008, I-Tech Education and Publishing, Vienna, Austria
This work presents an overview of current work applying reinforcement learning in medical image applications, presenting a detailed illustration of a particular use for lung nodules classification. The addressed application of reinforcement learning to solve the problem of lung nodules classification used the 3D geometric nodules characteristics to guide the classification. Even though the results are preliminary we may see that the obtained results are very encouraging, demonstrating that the reinforcement learning classifier using characteristics of the nodules’ geometry can effectively classify benign from malignant lung nodules based on CT images. On the other side, we may observe that this is a machine learning that is not commonly applied to medical images and this is an opportunity for more intensive investment in the research for this area. But some problems are well known in this application and must be more studied. We should research how to find out a way to shorten the training phase while maintaining the learning quality. And also must be improved the tests to generate more definitive results and to make possible the comparison with other classifiers.
ANALYSIS OF ENSEMBLE METHODS APPLIED TO LITHOLOGY CLASSIFICATION FROM WELL LOGS
Aug 2013 – 13th International Congress of The Brazilian Geophysical Society (SBGf)
This paper analyzes ensemble methods applied to automatic lithology classification. For this, we performed a comparison between single classifiers (Support Vector Machine and Multilayer Perceptron) and these classifiers with ensemble methods (Bagging and Boost).
MAPA DAS ESTRUTURAS DA MAMA USANDO KD-TREE+
Jun 2008 - Jornada de Informática do Maranhão (JIM)
Este artigo apresenta uma nova técnica para a segmentação de mamogramas. Com base em características da imagem radiográfica, aplicou-se o KD-Tree+ para divisão da imagem em vários grupos, nos quais podemos identificar algumas estruturas importantes como nódulos e calcificações. A técnica proposta foi aplicada em imagens da base de dados Mini-MIAS e os resultados foram avaliados através de inspeção médica.