Liquid State Machines (LSMs) exploit the power of recurrent spiking neural networks (SNNs) without training the SNN. Instead, a reservoir, or liquid, is randomly created which acts as a filter for a readout function. We develop three methods for iteratively refining a randomly generated liquid to create a more effective one. First, we apply Hebbian learning to LSMs by building the liquid with spike-time dependant plasticity (STDP) synapses. Second, we create an eligibility based reinforcement learning algorithm for synaptic development. Third, we apply principles of Hebbian learning and reinforcement learning to create a new algorithm called separation driven synaptic modification (SDSM). These three methods are compared across four artificial pattern recognition problems, generating only fifty liquids for each problem. Each of these algorithms shows overall improvements to LSMs with SDSM demonstrating the greatest improvement. SDSM is also shown to generalize well and outperforms traditional LSMs when presented with speech data obtained from the TIMIT dataset.

This paper introduces a new way of solving an edge user allocation problem. The problem is to be solved with a network of spiking neurons. This network should quickly and with low energy cost solve the optimization problem of allocating users to servers and minimizing the amount of servers hired to reduce the related hiring cost. The demonstrated method is a simulation of a method which could be implemented onto neuromorphic hardware. It is written in Python using the Brian2 spiking neural network simulator. The core of the method involves simulating an energy function through the use of circuit motifs. The dynamics of these circuit motifs mimic a search for the lowest energy point in an energy landscape, corresponding to a valid solution for the edge user allocation problem. The paper also shows the results of testing this network within the Brian2 environment.

The Chinese room argument (CRA) was first stated in 1980. Since then computer technologies have improved and today spiking neural networks (SNNs) are “arguably the only viable option if one wants to understand how the brain computes.” (Tavanei et.al. 2019: 47) SNNs differ in various important respects from the digital computers the CRA was directed against. The objective of the present work is to explore whether the CRA applies to SNNs. In the first chapter I am going to discuss computationalism, the Chinese room argument and give a brief overview over spiking neural networks. The second chapter is going to be considered with five important differences between SNNs and digital computers: (1) Massive parallelism, (2) subsymbolic computation, (3) machine learning, (4) analogue representation and (5) temporal encoding. I am going to finish by concluding that, besides minor limitations, the Chinese room argument can be applied to spiking neural networks.:1 Introduction 2 Theoretical background 2.I Strong AI: Computationalism 2.II The Chinese room argument 2.III Spiking neural networks 3 Applicability to spiking neural networks 3.I Massive parallelism 3.II Subsymbolic computation 3.III Machine learning 3.IV Analogue representation 3.V Temporal encoding 3.VI The Morse code room and its replies 3.VII Some more general considerations regarding hardware and software 4 Conclusion / Das Argument vom chinesischen Zimmer wurde erstmals 1980 veröffentlicht. Seit dieser Zeit hat sich die Computertechnologie stark weiterentwickelt und die heute viel beachteten gepulsten neuronalen Netze ähneln stark dem Aufbau und der Arbeitsweise biologischer Gehirne. Gepulste neuronale Netze unterscheiden sich in verschiedenen wichtigen Aspekten von den digitalen Computern, gegen die die CRA gerichtet war. Das Ziel der vorliegenden Arbeit ist es, zu untersuchen, ob das Argument vom chinesischen Zimmer auf gepulste neuronale Netze anwendbar ist. Im ersten Kapitel werde ich den Computer-Funktionalismus und das Argument des chinesischen Zimmers erörtern und einen kurzen Überblick über gepulste neuronale Netze geben. Das zweite Kapitel befasst sich mit fünf wichtigen Unterschieden zwischen gepulsten neuronalen Netzen und digitalen Computern: (1) Massive Parallelität, (2) subsymbolische Berechnung, (3) maschinelles Lernen, (4) analoge Darstellung und (5) zeitliche Kodierung. Ich werde schlussfolgern, dass das Argument des chinesischen Zimmers, abgesehen von geringfügigen Einschränkungen, auf gepulste neuronale Netze angewendet werden kann.:1 Introduction 2 Theoretical background 2.I Strong AI: Computationalism 2.II The Chinese room argument 2.III Spiking neural networks 3 Applicability to spiking neural networks 3.I Massive parallelism 3.II Subsymbolic computation 3.III Machine learning 3.IV Analogue representation 3.V Temporal encoding 3.VI The Morse code room and its replies 3.VII Some more general considerations regarding hardware and software 4 Conclusion

Generative models are broadly used in many subfields of DL. DNNs have recently developed a core approach to solving data-centric problems in image classification, translation, etc. The latest developments in parameterizing these models using DNNs and stochastic optimization algorithms have allowed scalable modeling of complex, high-dimensional data, including speech, text, and image. This dissertation proposal presents our state-the-art probabilistic bases and DL algorithms for generative models, including VAEs, GANs, and RNN-based encoder-decoder. The proposal also discusses application areas that may benefit from deep generative models in both NLP and computer vision. In NLP, we proposed an Arabic poetry generation model with extended phonetic and semantic embeddings (Phonetic CNN_subword embeddings). Extensive quantitative experiments using BLEU scores and Hamming distance show notable enhancements over strong baselines. Additionally, a comprehensive human evaluation confirms that the poems generated by our model outperform the base models in criteria including meaning, coherence, fluency, and poeticness. We proposed a generative video model using a hybrid VAE-GAN model in computer vision. Besides, we integrate two attentional mechanisms with GAN to get the essential regions of interest in a video, focused on enhancing the visual implementation of the human motion in the generated output. We have considered quantitative and qualitative experiments, including comparisons with other state-of-the-arts for evaluation. Our results indicate that our model enhances performance compared with other models and performs favorably under different quantitive metrics PSNR, SSIM, LPIPS, and FVD.Recently, mimicking biologically inspired learning in generative models based on SNNs has been shown their effectiveness in different applications. SNNs are the third generation of neural networks, in which neurons communicate through binary signals known as spikes. Since SNNs are more energy-efficient than DNNs. Moreover, DNN models have been vulnerable to small adversarial perturbations that cause misclassification of legitimate images. This dissertation shows the proposed ``VAE-Sleep'' that combines ideas from VAE and the sleep mechanism leveraging the advantages of deep and spiking neural networks (DNN--SNN).On top of that, we present ``Defense–VAE–Sleep'' that extended work of ``VAE-Sleep'' model used to purge adversarial perturbations from contaminated images. We demonstrate the benefit of sleep in improving the generalization performance of the traditional VAE when the testing data differ in specific ways even by a small amount from the training data. We conduct extensive experiments, including comparisons with the state–of–the–art on different datasets.

At a cocktail party, one can choose to scan the room for conversations of interest, attend to a specific conversation partner, switch between conversation partners, or not attend to anything at all. The ability of the normal-functioning auditory system to flexibly listen in complex acoustic scenes plays a central role in solving the cocktail party problem (CPP). In contrast, certain demographics (e.g., individuals with hearing impairment or older adults) are unable to solve the CPP, leading to psychological ailments and reduced quality of life. Since the normal auditory system still outperforms machines in solving the CPP, an effective solution may be found by mimicking the normal-functioning auditory system. Spatial hearing likely plays an important role in CPP-processing in the auditory system. This thesis details the development of a biologically based approach to the CPP by modeling specific neural mechanisms underlying spatial tuning in the auditory cortex. First, we modeled bottom-up, stimulus-driven mechanisms using a multi-layer network model of the auditory system. To convert spike trains from the model output into audible waveforms, we designed a novel reconstruction method based on the estimation of time-frequency masks. We showed that our reconstruction method produced sounds with significantly higher intelligibility and quality than previous reconstruction methods. We also evaluated the algorithm's performance using a psychoacoustic study, and found that it provided the same amount of benefit to normal-hearing listeners as a current state-of-the-art acoustic beamforming algorithm. Finally, we modeled top-down, attention driven mechanisms that allowed the network to flexibly operate in different regimes, e.g., monitor the acoustic scene, attend to a specific target, and switch between attended targets. The model explains previous experimental observations, and proposes candidate neural mechanisms underlying flexible listening in cocktail-party scenarios. The strategies proposed here would benefit hearing-assistive devices for CPP processing (e.g., hearing aids), where users would benefit from switching between various modes of listening in different social situations. / 2022-05-19T00:00:00Z

Abstract : Electroencephalography (EEG) is a non-invasive neuroimaging modality that was first introduced over 80 years ago. Surface EEG does not directly measure neuronal activity, and it is often assumed that it cannot provide indications on the underlying neuronal firing. However, recent studies based on invasive measurements in monkeys have shown that the coupling between two EEG frequency bands, namely the Gamma (25-45 Hz) and Delta (2-4 Hz) bands, is a good predictor of underlying mass-spiking activity. Specifically, when the Delta signal is in its trough and Gamma power is high, the probability of mass- firing of neurons is large. Here, we investigate this property in healthy human EEG acquired during resting-state. Using the interaction between Delta phase and Gamma power, we derived a modeled spike signal (MSS) from the recorded EEG. We found the power spectrum density (PSD) pattern of the MSS to be similar to that observed in animal studies. Specifically, between 1-10 Hz that the PSD deviates from a 1/[florin] trend and exhibits a small peak at about 2-3Hz. In addition, an inter-hemispheric correlation was found between the MSS of the different pairs of electrode in opposite hemispheres. Our results open the possibility of studying underlying neuronal output with non-invasive EEG. // Résumé : L'électroencéphalographie (EEG) est une modalité de neuro-imagerie non invasive qui a été introduite il y a plus de 80 ans. L’EEG de surface ne mesure pas directement l’activité neuronale et il est généralement supposé qu’elle ne donne pas d’indications sur la décharge neuronale sous-jacente. Cependant des études récentes ont montré à l’aide de mesures invasives que le couplage entre deux bandes de fréquences EEG, soit les bandes Gamma (25-45 Hz) et Delta (2-4 Hz), est un bon indicateur de l’activité neuronale de masse sous-jacente chez les singes. Plus précisément, lorsque le signal Delta est dans un creux (phase de π) et que la puissance dans le signal Gamma est élevée, la probabilité de décharge de masse des neurones est grande. Cette propriété est ici étudiée dans les signaux EEG d’humains sains en état de repos. En se basant sur l'interaction entre la phase du signal Delta et la puissance du signal Gamma, nous avons dérivé un modèle de l’activité neuronale de masse sous-jacente (modeled spike signal-MSS) obtenu à partir du signal l'EEG enregistrée. On trouve que la densité spectrale de puissance (power spectal density-PSD) du MSS est similaire à celle observée dans les études animales. Plus spécifiquement, entre 1-10 Hz la PSD s’écarte d’une tendance en 1 / [florin] et présente un pic de faible amplitude à environ 2-3Hz. En outre, une corrélation inter-hémisphérique a été observée entre les MSS de différentes paires d'électrodes positionnées sur les hémisphères opposés. Nos résultats ouvrent la possibilité d'étudier l’activité neuronale sous-jacente par EEG non-invasive.

Online monitoring applications requiring advanced pattern recognition capabilities implemented in resource-constrained wireless sensor systems are challenging to construct using standard digital computers. An interesting alternative solution is to use a low-power neuromorphic processor like the ROLLS, with subthreshold mixed analog/digital circuits and online learning capabilities that approximate the behavior of real neurons and synapses. This requires that the monitoring algorithm is implemented with spiking neural networks, which in principle are efficient computational models for tasks such as pattern recognition. In this work, I investigate how spiking neural networks can be used as a pre-processing and feature learning system in a condition monitoring application where the vibration of a machine with healthy and faulty rolling-element bearings is considered. Pattern recognition with spiking neural networks is investigated using simulations with Brian -- a Python-based open source toolbox -- and an implementation is developed for the ROLLS neuromorphic processor. I analyze the learned feature-response properties of individual neurons. When pre-processing the input signals with a neuromorphic cochlea known as the AER-EAR system, the ROLLS chip learns to classify the resulting spike patterns with a training error of less than 1 %, at a combined power consumption of approximately 30 mW. Thus, the neuromorphic hardware system can potentially be realized in a resource-constrained wireless sensor for online monitoring applications.However, further work is needed for testing and cross validation of the feature learning and pattern recognition networks.i

Cortical and subcortical microcircuits are continuously modified throughout life. Despite ongoing changes these networks stubbornly maintain their functions, which persist although destabilizing synaptic and nonsynaptic mechanisms should ostensibly propel them towards runaway excitation or quiescence. What dynamical phenomena exist to act together to balance such learning with information processing? What types of activity patterns do they underpin, and how do these patterns relate to our perceptual experiences? What enables learning and memory operations to occur despite such massive and constant neural reorganization? Progress towards answering many of these questions can be pursued through large-scale neuronal simulations.    In this thesis, a Hebbian learning rule for spiking neurons inspired by statistical inference is introduced. The spike-based version of the Bayesian Confidence Propagation Neural Network (BCPNN) learning rule involves changes in both synaptic strengths and intrinsic neuronal currents. The model is motivated by molecular cascades whose functional outcomes are mapped onto biological mechanisms such as Hebbian and homeostatic plasticity, neuromodulation, and intrinsic excitability. Temporally interacting memory traces enable spike-timing dependence, a stable learning regime that remains competitive, postsynaptic activity regulation, spike-based reinforcement learning and intrinsic graded persistent firing levels.    The thesis seeks to demonstrate how multiple interacting plasticity mechanisms can coordinate reinforcement, auto- and hetero-associative learning within large-scale, spiking, plastic neuronal networks. Spiking neural networks can represent information in the form of probability distributions, and a biophysical realization of Bayesian computation can help reconcile disparate experimental observations. / <p>QC 20170421</p>

Redes neurais pulsadas - redes que utilizam uma codificação temporal da informação - têm despontado como uma nova e promissora abordagem dentro do paradigma conexionista emergente da ciência cognitiva. Um desses novos modelos é a rede neural pulsada de base radial, capaz de armazenar informação nos tempos de atraso axonais dos neurônios e que comporta algoritmos explícitos de treinamento. A recente proposição de uma sistemática para a codificação temporal dos dados de entrada utilizando campos receptivos gaussianos tem apresentado interessantes resultados na tarefa do agrupamento de dados (clustering). Este trabalho propõe uma função para o aprendizado não supervisionado dessa rede, com o objetivo de simplificar a sistemática de calibração de alguns dos seus parâmetros-chave, aprimorando a convergência da rede neural pulsada no aprendizado baseado em instâncias. O desempenho desse modelo é avaliado na tarefa de classificação de padrões, particularmente na classificação de pixels em imagens coloridas no domínio da visão computacional. / Pulsed neural networks - networks that encode information in the timing of spikes - have been studied as a new and promising approach in the artificial neural networks paradigm, emergent from cognitive science. One of these new models is the pulsed neural network with radial basis function, a network able to store information in the axonal propagation delay of neurons. Recently, a new method for encoding input-data by population code using gaussian receptive fields has showed interesting results in the clustering task. The present work proposes a function for the unsupervised learning task in this network, which goal includes the simplification of the calibration of the network key parameters and the enhancement of the pulsed neural network convergence to instance based learning. The performance of this model is evaluated for pattern classification, particularly for the pixel colors classification task, in the computer vision domain.

Este trabalho trata sobre a transferência de frequência em neurônios de disparo, especificamente neurônios integra-e-dispara com escoamento e neurônios de Izhikevich. Através de análises matemáticas analíticas e sistemáticas simulações numéricas é obtida a função de ganho, a transferência de frequência estacionária e dinâmica dos neurônios utilizados, para diversos valores dos parâmetros do modelo. Desse modo, são realizados múltiplos ajustes às curvas obtidas, e os coeficientes estimados são apresentados. Com base em todos esses dados, são obtidas diversas características dessas relações de transferência de frequência, e como suas propriedades variam com relação aos principais parâmetros do modelo de neurônio e sinapse utilizados. Diversos resultados interessantes foram apresentados, incluindo evidências de que a função ganho do neurônio integra-e-dispara pode se comportar de modo bastante semelhante à função de ganho e transferência estacionária do neurônio de Izhikevich, dependendo dos parâmetros adotados; a divisão do plano de parâmetros do modelo integra-e-dispara de acordo com a linearidade da transferência de frequência dinâmica; o limiar da intensidade de corrente contínua e de frequência de spikes pré-sinápticos de um neurônio de Izhikevich é determinado apenas pelo parâmetro b, no intervalo de parâmetros usual; modelos de sinapses distintos tendem a não alterar a forma da transferência de frequência estacionária de um neurônio de Izhikevich. / This work is about the frequency transfer of spiking neurons, specifically integrate-and-fire neurons and Izhikevich neurons. Through analytical and systematic numerical simulations the gain function, the stationary and dynamic frequency transfer of the adopted neuron models, are obtained for several values of the model parameters. Thus, multiple fits are made to the curves obtained, and the estimated coefficients are presented. Based on all these data, several characteristics of the frequency transfer relations are obtained, and information is obtained about how their properties vary with respect the parameters of the adopted neuron and synapse model. Several interesting results have been presented, including evidences that the integrate-and-fire neuron\'s gain function can behave quite similarly to the Izhikevich neuron\'s stationary transfer and gain function, depending of the adopted parameters. We also obtained the division of the parameters plane of integrate-and-fire model according to the linearity of the dynamic frequency transfer. It was also verified that the thresholds of the presynaptic spikes\' current intensity and frequency of an Izhikevich neuron are determined only by the parameter b, in the usual parameter range. In addition, it was observed that the considered distinct synapses models tend not to depart from the stationary frequency transfer of an Izhikevich neuron.