arXiv subjects

Jean Rouat

Publications and source records attributed to Jean Rouat.

At least 19 recordsLinked to original sources

arXiv2026

Learning to See via Epiretinal Implant Stimulation in silico with Model-Based Deep Reinforcement Learning

Objective: Diseases such as age-related macular degeneration and retinitis pigmentosa cause the degradation of the photoreceptor layer. One approach to restore vision is to electrically stimulate the surviving retinal ganglion cells with a microelectrode array such as epiretinal implants. Epiretinal implants are known to generate visible anisotropic shapes elongated along the axon fascicles of neighboring retinal ganglion cells. Recent work has demonstrated that to obtain isotropic pixel-like shapes, it is possible to map axon fascicles and avoid stimulating them by inactivating electrodes or lowering stimulation current levels. Avoiding axon fascicle stimulation aims to remove brushstroke-like shapes in favor of a more reduced set of pixel-like shapes. Approach: In this study, we propose the use of isotropic and anisotropic shapes to render intelligible images on the retina of a virtual patient in a reinforcement learning environment named rlretina. The environment formalizes the task as using brushstrokes in a stroke-based rendering task. Main Results: We train a deep reinforcement learning agent that learns to assemble isotropic and anisotropic shapes to form an image. We investigate which error-based or perception-based metrics is adequate to reward the agent. The agent is trained in a model-based data generation fashion using the psychophysically validated axon map model to render images as perceived by different virtual patients. We show that the agent can generate more intelligible images compared to the naive method in different virtual patients. Significance: This work shares a new way to address epiretinal stimulation that constitutes a first step towards improving visual acuity in artificially-restored vision using anisotropic phosphenes.

cs.LG↗

arXiv2025

Adaptive Central Frequencies Locally Competitive Algorithm for Speech

Neuromorphic computing, inspired by nervous systems, revolutionizes information processing with its focus on efficiency and low power consumption. Using sparse coding, this paradigm enhances processing efficiency, which is crucial for edge devices with power constraints. The Locally Competitive Algorithm (LCA), adapted for audio with Gammatone and Gammachirp filter banks, provides an efficient sparse coding method for neuromorphic speech processing. Adaptive LCA (ALCA) further refines this method by dynamically adjusting modulation parameters, thereby improving reconstruction quality and sparsity. This paper introduces an enhanced ALCA version, the ALCA Central Frequency (ALCA-CF), which dynamically adapts both modulation parameters and central frequencies, optimizing the speech representation. Evaluations show that this approach improves reconstruction quality and sparsity while significantly reducing the power consumption of speech classification, without compromising classification accuracy, particularly on Intel's Loihi 2 neuromorphic chip.

cs.SD↗

arXiv2024

Maximizing Information in Neuron Populations for Neuromorphic Spike Encoding

Neuromorphic applications emulate the processing performed by the brain by using spikes as inputs instead of time-varying analog stimuli. Therefore, these time-varying stimuli have to be encoded into spikes, which can induce important information loss. To alleviate this loss, some studies use population coding strategies to encode more information using a population of neurons rather than just one neuron. However, configuring the encoding parameters of such a population is an open research question. This work proposes an approach based on maximizing the mutual information between the signal and the spikes in the population of neurons. The proposed algorithm is inspired by the information-theoretic framework of Partial Information Decomposition. Two applications are presented: blood pressure pulse wave classification, and neural action potential waveform classification. In both tasks, the data is encoded into spikes and the encoding parameters of the neuron populations are tuned to maximize the encoded information using the proposed algorithm. The spikes are then classified and the performance is measured using classification accuracy as a metric. Two key results are reported. Firstly, adding neurons to the population leads to an increase in both mutual information and classification accuracy beyond what could be accounted for by each neuron separately, showing the usefulness of population coding strategies. Secondly, the classification accuracy obtained with the tuned parameters is near-optimal and it closely follows the mutual information as more neurons are added to the population. Furthermore, the proposed approach significantly outperforms random parameter selection, showing the usefulness of the proposed approach. These results are reproduced in both applications.

cs.NE↗

arXiv2024

Efficient Sparse Coding with the Adaptive Locally Competitive Algorithm for Speech Classification

Researchers are exploring novel computational paradigms such as sparse coding and neuromorphic computing to bridge the efficiency gap between the human brain and conventional computers in complex tasks. A key area of focus is neuromorphic audio processing. While the Locally Competitive Algorithm has emerged as a promising solution for sparse coding, offering potential for real-time and low-power processing on neuromorphic hardware, its applications in neuromorphic speech classification have not been thoroughly studied. The Adaptive Locally Competitive Algorithm builds upon the Locally Competitive Algorithm by dynamically adjusting the modulation parameters of the filter bank to fine-tune the filters' sensitivity. This adaptability enhances lateral inhibition, improving reconstruction quality, sparsity, and convergence time, which is crucial for real-time applications. This paper demonstrates the potential of the Locally Competitive Algorithm and its adaptive variant as robust feature extractors for neuromorphic speech classification. Results show that the Locally Competitive Algorithm achieves better speech classification accuracy at the expense of higher power consumption compared to the LAUSCHER cochlea model used for benchmarking. On the other hand, the Adaptive Locally Competitive Algorithm mitigates this power consumption issue without compromising the accuracy. The dynamic power consumption is reduced to a range of 4 to 13 milliwatts on neuromorphic hardware, three orders of magnitude less than setups using Graphics Processing Units. These findings position the Adaptive Locally Competitive Algorithm as a compelling solution for efficient speech classification systems, promising substantial advancements in balancing speech classification accuracy and power efficiency.

eess.AS↗

arXiv2024

NAAQA: A Neural Architecture for Acoustic Question Answering

The goal of the Acoustic Question Answering (AQA) task is to answer a free-form text question about the content of an acoustic scene. It was inspired by the Visual Question Answering (VQA) task. In this paper, based on the previously introduced CLEAR dataset, we propose a new benchmark for AQA, namely CLEAR2, that emphasizes the specific challenges of acoustic inputs. These include handling of variable duration scenes, and scenes built with elementary sounds that differ between training and test set. We also introduce NAAQA, a neural architecture that leverages specific properties of acoustic inputs. The use of 1D convolutions in time and frequency to process 2D spectro-temporal representations of acoustic content shows promising results and enables reductions in model complexity. We show that time coordinate maps augment temporal localization capabilities which enhance performance of the network by ~17 percentage points. On the other hand, frequency coordinate maps have little influence on this task. NAAQA achieves 79.5% of accuracy on the AQA task with ~4 times fewer parameters than the previously explored VQA model. We evaluate the perfomance of NAAQA on an independent data set reconstructed from DAQA. We also test the addition of a MALiMo module in our model on both CLEAR2 and DAQA. We provide a detailed analysis of the results for the different question types. We release the code to produce CLEAR2 as well as NAAQA to foster research in this newly emerging machine learning task.

cs.CL↗

arXiv2024

Stabilizing Spiking Neuron Training

Stability arguments are often used to prevent learning algorithms from having ever increasing activity and weights that hinder generalization. However, stability conditions can clash with the sparsity required to augment the energy efficiency of spiking neurons. Nonetheless it can also provide solutions. In fact, spiking Neuromorphic Computing uses binary activity to improve Artificial Intelligence energy efficiency. However, its non-smoothness requires approximate gradients, known as Surrogate Gradients (SG), to close the performance gap with Deep Learning. Several SG have been proposed in the literature, but it remains unclear how to determine the best SG for a given task and network. Thus, we aim at theoretically define the best SG, through stability arguments, to reduce the need for grid search. In fact, we show that more complex tasks and networks need more careful choice of SG, even if overall the derivative of the fast sigmoid tends to outperform the other, for a wide range of learning rates. We therefore design a stability based theoretical method to choose initialization and SG shape before training on the most common spiking neuron, the Leaky Integrate and Fire (LIF). Since our stability method suggests the use of high firing rates at initialization, which is non-standard in the neuromorphic literature, we show that high initial firing rates, combined with a sparsity encouraging loss term introduced gradually, can lead to better generalization, depending on the SG shape. Our stability based theoretical solution, finds a SG and initialization that experimentally result in improved accuracy. We show how it can be used to reduce the need of extensive grid-search of dampening, sharpness and tail-fatness of the SG. We also show that our stability concepts can be extended to be applicable on different LIF variants, such as DECOLLE and fluctuations-driven initializations.

cs.NE↗

arXiv2024

Stabilizing RNN Gradients through Pre-training

Numerous theories of learning propose to prevent the gradient from exponential growth with depth or time, to stabilize and improve training. Typically, these analyses are conducted on feed-forward fully-connected neural networks or simple single-layer recurrent neural networks, given their mathematical tractability. In contrast, this study demonstrates that pre-training the network to local stability can be effective whenever the architectures are too complex for an analytical initialization. Furthermore, we extend known stability theories to encompass a broader family of deep recurrent networks, requiring minimal assumptions on data and parameter distribution, a theory we call the Local Stability Condition (LSC). Our investigation reveals that the classical Glorot, He, and Orthogonal initialization schemes satisfy the LSC when applied to feed-forward fully-connected neural networks. However, analysing deep recurrent networks, we identify a new additive source of exponential explosion that emerges from counting gradient paths in a rectangular grid in depth and time. We propose a new approach to mitigate this issue, that consists on giving a weight of a half to the time and depth contributions to the gradient, instead of the classical weight of one. Our empirical results confirm that pre-training both feed-forward and recurrent networks, for differentiable, neuromorphic and state-space models to fulfill the LSC, often results in improved final performance. This study contributes to the field by providing a means to stabilize networks of any complexity. Our approach can be implemented as an additional step before pre-training on large augmented datasets, and as an alternative to finding stable initializations analytically.

cs.LG↗

arXiv2023

The connectivity degree controls the difficulty of RBN reservoir design

Reservoir Computing (RC) is a paradigm in artificial intelligence where a recurrent neural network (RNN) is used to process temporal data, leveraging the inherent dynamical properties of the reservoir to perform complex computations. In the realm of RC, the excitatory-inhibitory balance b has been shown to be pivotal for driving the dynamics and performance of Echo State Networks (ESN) and, more recently, Random Boolean Network (RBN). However, the relationship between b and other parameters of the network is still poorly understood. This article explores how the interplay of the balance b, the connectivity degree K (i.e., the number of synapses per neuron) and the size of the network (i.e., the number of neurons N) influences the dynamics and performance (memory and prediction) of an RBN reservoir. Our findings reveal that K and b are strongly tied in optimal reservoirs. Reservoirs with high K have two optimal balances, one for globally inhibitory networks (b<0), and the other one for excitatory networks (b>0). Both show asymmetric performances about a zero balance. In contrast, for moderate K, the optimal value being K=4, best reservoirs are obtained when excitation and inhibition almost, but not exactly, balance each other. For almost all K, the influence of the size is such that increasing N leads to better performance, even with large values of N. Our investigation provides clear directions to generate optimal reservoirs or reservoirs with constraints on size or connectivity.

nlin.AO↗

arXiv2023

Expanding memory in recurrent spiking networks

Recurrent spiking neural networks (RSNNs) are notoriously difficult to train because of the vanishing gradient problem that is enhanced by the binary nature of the spikes. In this paper, we review the ability of the current state-of-the-art RSNNs to solve long-term memory tasks, and show that they have strong constraints both in performance, and for their implementation on hardware analog neuromorphic processors. We present a novel spiking neural network that circumvents these limitations. Our biologically inspired neural network uses synaptic delays, branching factor regularization and a novel surrogate derivative for the spiking function. The proposed network proves to be more successful in using the recurrent connections on memory tasks.

cs.NE↗

arXiv2023

Excitatory/Inhibitory Balance Emerges as a Key Factor for RBN Performance, Overriding Attractor Dynamics

Reservoir computing provides a time and cost-efficient alternative to traditional learning methods.Critical regimes, known as the "edge of chaos," have been found to optimize computational performance in binary neural networks. However, little attention has been devoted to studying reservoir-to-reservoir variability when investigating the link between connectivity, dynamics, and performance. As physical reservoir computers become more prevalent, developing a systematic approach to network design is crucial. In this article, we examine Random Boolean Networks (RBNs) and demonstrate that specific distribution parameters can lead to diverse dynamics near critical points. We identify distinct dynamical attractors and quantify their statistics, revealing that most reservoirs possess a dominant attractor. We then evaluate performance in two challenging tasks, memorization and prediction, and find that a positive excitatory balance produces a critical point with higher memory performance. In comparison, a negative inhibitory balance delivers another critical point with better prediction performance. Interestingly, we show that the intrinsic attractor dynamics have little influence on performance in either case.

q-bio.NC↗

arXiv2023

Evaluation of short range depth sonifications for visual-to-auditory sensory substitution

Visual to auditory sensory substitution devices convert visual information into sound and can provide valuable assistance for blind people. Recent iterations of these devices rely on depth sensors. Rules for converting depth into sound (i.e. the sonifications) are often designed arbitrarily, with no strong evidence for choosing one over another. The purpose of this work is to compare and understand the effectiveness of five depth sonifications in order to assist the design process of future visual to auditory systems for blind people which rely on depth sensors. The frequency, amplitude and reverberation of the sound as well as the repetition rate of short high-pitched sounds and the signal-to-noise ratio of a mixture between pure sound and noise are studied. We conducted positioning experiments with twenty-eight sighted blindfolded participants. Stage 1 incorporates learning phases followed by depth estimation tasks. Stage 2 adds the additional challenge of azimuth estimation to the first stage's protocol. Stage 3 tests learning retention by incorporating a 10-minute break before re-testing depth estimation. The best depth estimates in stage 1 were obtained with the sound frequency and the repetition rate of beeps. In stage 2, the beep repetition rate yielded the best depth estimation and no significant difference was observed for the azimuth estimation. Results of stage 3 showed that the beep repetition rate was the easiest sonification to memorize. Based on statistical analysis of the results, we discuss the effectiveness of each sonification and compare with other studies that encode depth into sounds. Finally we provide recommendations for the design of depth encoding.

cs.HC↗

arXiv2023

Evaluation of Neuromorphic Spike Encoding of Sound Using Information Theory

The problem of spike encoding of sound consists in transforming a sound waveform into spikes. It is of interest in many domains, including the development of audio-based spiking neural networks, where it is the first and most crucial stage of processing. Many algorithms have been proposed to perform spike encoding of sound. However, a systematic approach to quantitatively evaluate their performance is currently lacking. We propose the use of an information-theoretic framework to solve this problem. Specifically, we evaluate the coding efficiency of four spike encoding algorithms on two coding tasks that consist of coding the fundamental characteristics of sound: frequency and amplitude. The algorithms investigated are: Independent Spike Coding, Send-on-Delta coding, Ben's Spiker Algorithm, and Leaky Integrate-and-Fire coding. Using the tools of information theory, we estimate the information that the spikes carry on relevant aspects of an input stimulus. We find disparities in the coding efficiencies of the algorithms, where Leaky Integrate-and-Fire coding performs best. The information-theoretic analysis of their performance on these coding tasks provides insight on the encoding of richer and more complex sound stimuli.

cs.SD↗

arXiv2022

Voltage-Dependent Synaptic Plasticity (VDSP): Unsupervised probabilistic Hebbian plasticity rule based on neurons membrane potential

This study proposes voltage-dependent-synaptic plasticity (VDSP), a novel brain-inspired unsupervised local learning rule for the online implementation of Hebb's plasticity mechanism on neuromorphic hardware. The proposed VDSP learning rule updates the synaptic conductance on the spike of the postsynaptic neuron only, which reduces by a factor of two the number of updates with respect to standard spike-timing-dependent plasticity (STDP). This update is dependent on the membrane potential of the presynaptic neuron, which is readily available as part of neuron implementation and hence does not require additional memory for storage. Moreover, the update is also regularized on synaptic weight and prevents explosion or vanishing of weights on repeated stimulation. Rigorous mathematical analysis is performed to draw an equivalence between VDSP and STDP. To validate the system-level performance of VDSP, we train a single-layer spiking neural network (SNN) for the recognition of handwritten digits. We report 85.01 $ \pm $ 0.76% (Mean $ \pm $ S.D.) accuracy for a network of 100 output neurons on the MNIST dataset. The performance improves when scaling the network size (89.93 $ \pm $ 0.41% for 400 output neurons, 90.56 $ \pm $ 0.27 for 500 neurons), which validates the applicability of the proposed learning rule for spatial pattern recognition tasks. Future work will consider more complicated tasks. Interestingly, the learning rule better adapts than STDP to the frequency of input signal and does not require hand-tuning of hyperparameters

cs.NE↗

arXiv2022

Efficient spike encoding algorithms for neuromorphic speech recognition

Spiking Neural Networks (SNN) are known to be very effective for neuromorphic processor implementations, achieving orders of magnitude improvements in energy efficiency and computational latency over traditional deep learning approaches. Comparable algorithmic performance was recently made possible as well with the adaptation of supervised training algorithms to the context of SNN. However, information including audio, video, and other sensor-derived data are typically encoded as real-valued signals that are not well-suited to SNN, preventing the network from leveraging spike timing information. Efficient encoding from real-valued signals to spikes is therefore critical and significantly impacts the performance of the overall system. To efficiently encode signals into spikes, both the preservation of information relevant to the task at hand as well as the density of the encoded spikes must be considered. In this paper, we study four spike encoding methods in the context of a speaker independent digit classification system: Send on Delta, Time to First Spike, Leaky Integrate and Fire Neuron and Bens Spiker Algorithm. We first show that all encoding methods yield higher classification accuracy using significantly fewer spikes when encoding a bio-inspired cochleagram as opposed to a traditional short-time Fourier transform. We then show that two Send On Delta variants result in classification results comparable with a state of the art deep convolutional neural network baseline, while simultaneously reducing the encoded bit rate. Finally, we show that several encoding methods result in improved performance over the conventional deep learning baseline in certain cases, further demonstrating the power of spike encoding algorithms in the encoding of real-valued signals and that neuromorphic implementation has the potential to outperform state of the art techniques.

cs.NE↗

arXiv2022

Sonified distance in sensory substitution does not always improve localization: comparison with a 2D and 3D handheld device

Early visual to auditory substitution devices encode 2D monocular images into sounds while more recent devices use distance information from 3D sensors. This study assesses whether the addition of sound-encoded distance in recent systems helps to convey the "where" information. This is important to the design of new sensory substitution devices. We conducted experiments for object localization and navigation tasks with a handheld visual to audio substitution system. It comprises 2D and 3D modes. Both encode in real-time the position of objects in images captured by a camera. The 3D mode encodes in addition the distance between the system and the object. Experiments have been conducted with 16 blindfolded sighted participants. For the localization, participants were quicker to understand the scene with the 3D mode that encodes distances. On the other hand, with the 2D only mode, they were able to compensate for the lack of distance encoding after a small training. For the navigation, participants were as good with the 2D only mode than with the 3D mode encoding distance.

cs.HC↗

arXiv2022

Perceptive, non-linear Speech Processing and Spiking Neural Networks

Source separation and speech recognition are very difficult in the context of noisy and corrupted speech. Most conventional techniques need huge databases to estimate speech (or noise) density probabilities to perform separation or recognition. We discuss the potential of perceptive speech analysis and processing in combination with biologically plausible neural network processors. We illustrate the potential of such non-linear processing of speech on a source separation system inspired by an Auditory Scene Analysis paradigm. We also discuss a potential application in speech recognition.

cs.SD↗

arXiv2021

Adaptive Approach For Sparse Representations Using The Locally Competitive Algorithm For Audio

Gammachirp filterbank has been used to approximate the cochlea in sparse coding algorithms. An oriented grid search optimization was applied to adapt the gammachirp's parameters and improve the Matching Pursuit (MP) algorithm's sparsity along with the reconstruction quality. However, this combination of a greedy algorithm with a grid search at each iteration is computationally demanding and not suitable for real-time applications. This paper presents an adaptive approach to optimize the gammachirp's parameters but in the context of the Locally Competitive Algorithm (LCA) that requires much fewer computations than MP. The proposed method consists of taking advantage of the LCA's neural architecture to automatically adapt the gammachirp's filterbank using the backpropagation algorithm. Results demonstrate an improvement in the LCA's performance with our approach in terms of sparsity, reconstruction quality, and convergence time. This approach can yield a significant advantage over existing approaches for real-time applications.

cs.SD↗

arXiv2021

Signals to Spikes for Neuromorphic Regulated Reservoir Computing and EMG Hand Gesture Recognition

Surface electromyogram (sEMG) signals result from muscle movement and hence they are an ideal candidate for benchmarking event-driven sensing and computing. We propose a simple yet novel approach for optimizing the spike encoding algorithm's hyper-parameters inspired by the readout layer concept in reservoir computing. Using a simple machine learning algorithm after spike encoding, we report performance higher than the state-of-the-art spiking neural networks on two open-source datasets for hand gesture recognition. The spike encoded data is processed through a spiking reservoir with a biologically inspired topology and neuron model. When trained with the unsupervised activity regulation CRITICAL algorithm to operate at the edge of chaos, the reservoir yields better performance than state-of-the-art convolutional neural networks. The reservoir performance with regulated activity was found to be 89.72% for the Roshambo EMG dataset and 70.6% for the EMG subset of sensor fusion dataset. Therefore, the biologically-inspired computing paradigm, which is known for being power efficient, also proves to have a great potential when compared with conventional AI algorithms.

eess.SP↗