Papers by garibaldi pineda garcía
2016 IEEE Symposium Series on Computational Intelligence (SSCI), 2016
Vision is one of our most important senses, a vast amount of information is perceived through our... more Vision is one of our most important senses, a vast amount of information is perceived through our eyes. Neuroscientists have performed many studies using vision as input to their experiments. Computational neuroscientists have typically used a brightness-to-rate encoding to use images as spike-based visual sources for its natural mapping. Recently, neuromorphic Dynamic Vision Sensors (DVSs) were developed and, while they have excellent capabilities, they remain scarce and relatively expensive. We propose a visual input system inspired by the behaviour of a DVS but using a conventional digital camera as a sensor and a PC to encode the images. By using readily-available components, we believe most scientists would have access to a realistic spiking visual input source. While our primary goal is to provide systems with a live real-time input, we have also been successful in transcoding well established image and video databases into spike train representations. Our main contribution is a DVS emulator framework which can be extended, as we demonstrate by adding local inhibitory behaviour, adaptive thresholds and spike-timing encoding.
arXiv (Cornell University), Dec 20, 2022
Models of sensory processing and learning in the cortex need to efficiently assign credit to syna... more Models of sensory processing and learning in the cortex need to efficiently assign credit to synapses in all areas. In deep learning, a known solution is error backpropagation, which however requires biologically implausible weight transport from feed-forward to feedback paths. We introduce Phaseless Alignment Learning (PAL), a bio-plausible method to learn efficient feedback weights in layered cortical hierarchies. This is achieved by exploiting the noise naturally found in biophysical systems as an additional carrier of information. In our dynamical system, all weights are learned simultaneously with always-on plasticity and using only information locally available to the synapses. Our method is completely phase-free (no forward and backward passes or phased learning) and allows for efficient error propagation across multi-layer cortical hierarchies, while maintaining biologically plausible signal transport and learning. Our method is applicable to a wide class of models and improves on previously known biologically plausible ways of credit assignment: compared to random synaptic feedback, it can solve complex tasks with less neurons and learn more useful latent representations. We demonstrate this on various classification tasks using a cortical microcircuit model with prospective coding.
2013 IEEE International Autumn Meeting on Power Electronics and Computing (ROPEC), 2013
Computer hardware is becoming ubiquitous, from kitchen appliances to research machines to cell ph... more Computer hardware is becoming ubiquitous, from kitchen appliances to research machines to cell phones. This availability of computing capable circuits has driven its cost down, thus making several applications possible on commercial hardware. Computer vision is one of such tasks, it is now possible to create computer based helping devices for the visually impaired. In recent years there has been an explotion of computer vision hardware for many applications, 3D scanners in particular which are used to interact with regular operative systems or games. This paper presents an apparatus which uses one of such devices (Kinect) developed by Microsoft that senses the world in 3D. This information is then processed with a low-power computer called the BeagleBone whose output will be sent to the user via mechanical pulses.
SpiNNaker: A Spiking Neural Network Architecture, 2020
This paper focuses on how artificial neural networks do their learning. This was done by going th... more This paper focuses on how artificial neural networks do their learning. This was done by going through an in-depth literature review focusing on the definition of learning, learning paradigms, learning rules and algorithms. A number of learning paradigms were identified and explained, which included supervised learning, unsupervised learning, hybrid learning and reinforcement learning. Various learning rules and their associated algorithms were also explained and illustrated, which includes
Frontiers in neuroscience, 2018
SpiNNaker is a digital neuromorphic architecture, designed specifically for the low power simulat... more SpiNNaker is a digital neuromorphic architecture, designed specifically for the low power simulation of large-scale spiking neural networks at speeds close to biological real-time. Unlike other neuromorphic systems, SpiNNaker allows users to develop their own neuron and synapse models as well as specify arbitrary connectivity. As a result SpiNNaker has proved to be a powerful tool for studying different neuron models as well as synaptic plasticity-believed to be one of the main mechanisms behind learning and memory in the brain. A number of Spike-Timing-Dependent-Plasticity(STDP) rules have already been implemented on SpiNNaker and have been shown to be capable of solving various learning tasks in real-time. However, while STDP is an important biological theory of learning, it is a form of Hebbian or unsupervised learning and therefore does not explain behaviors that depend on feedback from the environment. Instead, learning rules based on neuromodulated STDP (three-factor learning ...
2014 IEEE International Autumn Meeting on Power, Electronics and Computing (ROPEC), 2014
This work proposes a qualitative representation for bifurcation diagrams with two varying paramet... more This work proposes a qualitative representation for bifurcation diagrams with two varying parameters. We start from a previous defined representation for a single parameter bifurcation diagram and add the neccesary definitions/constraints to extend it. In order to give sense to our proposed representation for bifurcation diagrams an algoritmh to perform qualitative simulations is given; this algorithm is based on behavioural inference rules for a system going through dynamic events. The developed algorithm is capable to generated all possible qualitative behaviours under a user defined sequence of qualitative events and parameters changes.
Interface focus, Jan 6, 2018
State-of-the-art computer vision systems use frame-based cameras that sample the visual scene as ... more State-of-the-art computer vision systems use frame-based cameras that sample the visual scene as a series of high-resolution images. These are then processed using convolutional neural networks using neurons with continuous outputs. Biological vision systems use a quite different approach, where the eyes (cameras) sample the visual scene continuously, often with a non-uniform resolution, and generate neural spike events in response to changes in the scene. The resulting spatio-temporal patterns of events are then processed through networks of spiking neurons. Such event-based processing offers advantages in terms of focusing constrained resources on the most salient features of the perceived scene, and those advantages should also accrue to engineered vision systems based upon similar principles. Event-based vision sensors, and event-based processing exemplified by the SpiNNaker (Spiking Neural Network Architecture) machine, can be used to model the biological vision pathway at vari...
SpiNNaker: A Spiking Neural Network Architecture
Se presenta un método alternativo basado en Polinomios de Chebyshev para análisis de redes eléctr... more Se presenta un método alternativo basado en Polinomios de Chebyshev para análisis de redes eléctricas cuya dinámica se puede modelar por ecuaciones de estado. La metodología toma las ventajas de las propiedades operacionales disponibles para la mayoría de las series ortogonales, como son, matriz de integración, derivación, producto y de coeficientes. Un hecho importante de esta aproximación es la obtención de una expresión analítica para el análisis del estado transitorio. Para mostrar la validación del método se usa una red eléctrica. Los resultados obtenidos son muy satisfactorios.
I want to thank the people of México whose contributions allowed the National Council for Science... more I want to thank the people of México whose contributions allowed the National Council for Science and Technology (CONACyT) to sponsor my PhD studies. I would like to show my deepest appreciation to my colleagues John V. Woods, Simon Davidson, Michael Hopkins, Luis Plana and Jim Garside for the diverse topics we had the chance to discuss. Their knowledge has nurtured my understanding of a broad range of topics. My eternal gratitude goes to James Knight, Alan Stokes, Andrew Rowley and Oliver Rhodes for their help in disentangling my view of the SpiNNaker software stack. Many thanks to my lab partners Qian Liu, Petruţ Bogdan, Mantas Mikaitis, Robert James, Patrick Camilleri and James Knight for the talks and collaborations which made my research possible. The path towards the PhD would have not been bearable without the presence of my friends. I would like to thank
This work presents an analysis of four regression systems. Two of them are statistical: the widel... more This work presents an analysis of four regression systems. Two of them are statistical: the widely used Auto-regressive Integrated Moving Average (ARIMA) and the state-of-the-art Facebook Prophet. From the deep learning school, a Long Short-Term Memory (LSTM) Recurrent Neural Network (RNN) is also evaluated. We finish our quartet with a fine-tuned Nearest Neighbor model. The study is carried out over seventeen benchmarks; fifteen coming from M4-Competition and two more power systems time series, i.e., electricity demand and hydropower generation. For all the models, the regression systems are fitted and optimized to minimize user intervention. The results show that deep learning models obtained the best performance; nonetheless, the performance difference is not statistically significant with the rest of the systems tested.
Frontiers in neuroscience, 2016
Today, increasing attention is being paid to research into spike-based neural computation both to... more Today, increasing attention is being paid to research into spike-based neural computation both to gain a better understanding of the brain and to explore biologically-inspired computation. Within this field, the primate visual pathway and its hierarchical organization have been extensively studied. Spiking Neural Networks (SNNs), inspired by the understanding of observed biological structure and function, have been successfully applied to visual recognition and classification tasks. In addition, implementations on neuromorphic hardware have enabled large-scale networks to run in (or even faster than) real time, making spike-based neural vision processing accessible on mobile robots. Neuromorphic sensors such as silicon retinas are able to feed such mobile systems with real-time visual stimuli. A new set of vision benchmarks for spike-based neural processing are now needed to measure progress quantitatively within this rapidly advancing field. We propose that a large dataset of spike...
ABSTRACT Developing web based information services has become a complex problem, in this paper we... more ABSTRACT Developing web based information services has become a complex problem, in this paper we present an approach to ease web development using the Django framework and the Google AJAX library(closure). Django allows developers to concentrate in the crucial tasks of web design since it provides high level abstractions of common web development patterns, shortcuts for frequent programming tasks, and clear conventions for how to solve problems. Closure library adds the dynamic interface part to web applications, it helps to increase the efficiency of data transmission, since it enables programmers to update web interfaces and content in a very flexible way. keywords: Django, web development, AJAX, Google Closure.
Uploads
Papers by garibaldi pineda garcía