arXiv subjects

Marcel van Gerven

Publications and source records attributed to Marcel van Gerven.

At least 55 records · Page 3Linked to original sources

arXiv2025

Leveraging Spatial Cues from Cochlear Implant Microphones to Efficiently Enhance Speech Separation in Real-World Listening Scenes

Speech separation approaches for single-channel, dry speech mixtures have significantly improved. However, real-world spatial and reverberant acoustic environments remain challenging, limiting the effectiveness of these approaches for assistive hearing devices like cochlear implants (CIs). To address this, we quantify the impact of real-world acoustic scenes on speech separation and explore how spatial cues can enhance separation quality efficiently. We analyze performance based on implicit spatial cues (inherent in the acoustic input and learned by the model) and explicit spatial cues (manually calculated spatial features added as auxiliary inputs). Our findings show that spatial cues (both implicit and explicit) improve separation for mixtures with spatially separated and nearby talkers. Furthermore, spatial cues enhance separation when spectral cues are ambiguous, such as when voices are similar. Explicit spatial cues are particularly beneficial when implicit spatial cues are weak. For instance, single CI microphone recordings provide weaker implicit spatial cues than bilateral CIs, but even single CIs benefit from explicit cues. These results emphasize the importance of training models on real-world data to improve generalizability in everyday listening scenarios. Additionally, our statistical analyses offer insights into how data properties influence model performance, supporting the development of efficient speech separation approaches for CIs and other assistive devices in real-world settings.

cs.SD↗

arXiv2025

Neural encoding with affine feature response transforms

Current linearizing encoding models that predict neural responses to sensory input typically neglect neuroscience-inspired constraints that could enhance model efficiency and interpretability. To address this, we propose a new method called affine feature response transform (AFRT), which exploits the brain's retinotopic organization. Applying AFRT to encode multi-unit activity in areas V1, V4, and IT of the macaque brain, we demonstrate that AFRT reduces redundant computations and enhances the performance of current linearizing encoding models by segmenting each neuron's receptive field into an affine retinal transform, followed by a localized feature response. Remarkably, by factorizing receptive fields into a sequential affine component with three interpretable parameters (for shifting and scaling) and response components with a small number of feature weights per response, AFRT achieves encoding with orders of magnitude fewer parameters compared to unstructured models. We show that the retinal transform of each neuron's encoding agrees well with the brain's receptive field. Together, these findings suggest that this new subset within spatial transformer network can be instrumental in neural encoding models of naturalistic stimuli.

q-bio.NC↗

arXiv2025

Inverse receptive field attention for naturalistic image reconstruction from the brain

Visual perception in the brain largely depends on the organization of neuronal receptive fields. Although extensive research has delineated the coding principles of receptive fields, most studies have been constrained by their foundational assumptions. Moreover, while machine learning has successfully been used to reconstruct images from brain data, this approach faces significant challenges, including inherent feature biases in the model and the complexities of brain structure and function. In this study, we introduce an inverse receptive field attention (IRFA) model, designed to reconstruct naturalistic images from neurophysiological data in an end-to-end fashion. This approach aims to elucidate the tuning properties and representational transformations within the visual cortex. The IRFA model incorporates an attention mechanism that determines the inverse receptive field for each pixel, weighting neuronal responses across the visual field and feature spaces. This method allows for an examination of the dynamics of neuronal representations across stimuli in both spatial and feature dimensions. Our results show highly accurate reconstructions of naturalistic data, independent of pre-trained models. Notably, IRF models trained on macaque V1, V4, and IT regions yield remarkably consistent spatial receptive fields across different stimuli, while the features to which neuronal representations are selective exhibit significant variation. Additionally, we propose a data-driven method to explore representational clustering within various visual areas, further providing testable hypotheses.

q-bio.NC↗

arXiv2024

Ornstein-Uhlenbeck Adaptation as a Mechanism for Learning in Brains and Machines

Learning is a fundamental property of intelligent systems, observed across biological organisms and engineered systems. While modern intelligent systems typically rely on gradient descent for learning, the need for exact gradients and complex information flow makes its implementation in biological and neuromorphic systems challenging. This has motivated the exploration of alternative learning mechanisms that can operate locally and do not rely on exact gradients. In this work, we introduce a novel approach that leverages noise in the parameters of the system and global reinforcement signals. Using an Ornstein-Uhlenbeck process with adaptive dynamics, our method balances exploration and exploitation during learning, driven by deviations from error predictions, akin to reward prediction error. Operating in continuous time, Orstein-Uhlenbeck adaptation (OUA) is proposed as a general mechanism for learning dynamic, time-evolving environments. We validate our approach across diverse tasks, including supervised learning and reinforcement learning in feedforward and recurrent systems. Additionally, we demonstrate that it can perform meta-learning, adjusting hyper-parameters autonomously. Our results indicate that OUA provides a viable alternative to traditional gradient-based methods, with potential applications in neuromorphic computing. It also hints at a possible mechanism for noise-driven learning in the brain, where stochastic neurotransmitter release may guide synaptic adjustments.

cs.LG↗

arXiv2024

Fractional Order Distributed Optimization

Distributed optimization is fundamental to modern machine learning applications like federated learning, but existing methods often struggle with ill-conditioned problems and face stability-versus-speed tradeoffs. We introduce fractional order distributed optimization (FrODO); a theoretically-grounded framework that incorporates fractional-order memory terms to enhance convergence properties in challenging optimization landscapes. Our approach achieves provable linear convergence for any strongly connected network. Through empirical validation, our results suggest that FrODO achieves up to 4 times faster convergence versus baselines on ill-conditioned problems and 2-3 times speedup in federated neural network training, while maintaining stability and theoretical guarantees.

cs.LG↗

arXiv2024

Modeling and Detection of Critical Slowing Down in Epileptic Dynamics

Epilepsy is a common neurological disorder characterized by abrupt seizures. Although seizures may appear random, they are often preceded by early warning signs in neural signals, notably, critical slowing down, a phenomenon in which the system's recovery rate from perturbations declines when it approaches a critical point. Detecting these markers could enable preventive therapies. This paper introduces a multi-stable slow-fast system to capture critical slowing down in epileptic dynamics. We construct regions of attraction for stable states, shedding light on how dynamic bifurcations drive pathological oscillations. We derive the recovery rate after perturbations to formalize critical slowing down. A novel algorithm for detecting precursors to ictal transitions is presented, along with a proof-of-concept event-based feedback control strategy to prevent impending pathological oscillations. Numerical studies are conducted to validate our theoretical findings.

eess.SY↗

arXiv2024

Gradient-Free Training of Recurrent Neural Networks using Random Perturbations

Recurrent neural networks (RNNs) hold immense potential for computations due to their Turing completeness and sequential processing capabilities, yet existing methods for their training encounter efficiency challenges. Backpropagation through time (BPTT), the prevailing method, extends the backpropagation (BP) algorithm by unrolling the RNN over time. However, this approach suffers from significant drawbacks, including the need to interleave forward and backward phases and store exact gradient information. Furthermore, BPTT has been shown to struggle to propagate gradient information for long sequences, leading to vanishing gradients. An alternative strategy to using gradient-based methods like BPTT involves stochastically approximating gradients through perturbation-based methods. This learning approach is exceptionally simple, necessitating only forward passes in the network and a global reinforcement signal as feedback. Despite its simplicity, the random nature of its updates typically leads to inefficient optimization, limiting its effectiveness in training neural networks. In this study, we present a new approach to perturbation-based learning in RNNs whose performance is competitive with BPTT, while maintaining the inherent advantages over gradient-based learning. To this end, we extend the recently introduced activity-based node perturbation (ANP) method to operate in the time domain, leading to more efficient learning and generalization. We subsequently conduct a range of experiments to validate our approach. Our results show similar performance, convergence time and scalability compared to BPTT, strongly outperforming standard node and weight perturbation methods. These findings suggest that perturbation-based learning methods offer a versatile alternative to gradient-based methods for training RNNs which can be ideally suited for neuromorphic computing applications

cs.LG↗

arXiv2024

Energy-Efficient Spiking Recurrent Neural Network for Gesture Recognition on Embedded GPUs

Implementing AI algorithms on event-based embedded devices enables real-time processing of data, minimizes latency, and enhances power efficiency in edge computing. This research explores the deployment of a spiking recurrent neural network (SRNN) with liquid time constant neurons for gesture recognition. We focus on the energy efficiency and computational efficacy of NVIDIA Jetson Nano embedded GPU platforms. The embedded GPU showcases a 14-fold increase in power efficiency relative to a conventional GPU, making a compelling argument for its use in energy-constrained applications. The study's empirical findings also highlight that batch processing significantly boosts frame rates across various batch sizes while maintaining accuracy levels well above the baseline. These insights validate the SRNN with liquid time constant neurons as a robust model for interpreting temporal-spatial data in gesture recognition, striking a critical balance between processing speed and power frugality.

cs.LG↗

arXiv2024

Analytical Characterization of Epileptic Dynamics in a Bistable System

Epilepsy is one of the most common neurological disorders globally, affecting millions of individuals. Despite significant advancements, the precise mechanisms underlying this condition remain largely unknown, making accurately predicting and preventing epileptic seizures challenging. In this paper, we employ a bistable model, where a stable equilibrium and a stable limit cycle coexist, to describe epileptic dynamics. The equilibrium captures normal steady-state neural activity, while the stable limit cycle signifies seizure-like oscillations. The noise-driven switch from the equilibrium to the limit cycle characterizes the onset of seizures. The differences in the regions of attraction of these two stable states distinguish epileptic brain dynamics from healthy ones. We analytically construct the regions of attraction for both states. Further, using the notion of input-to-state stability, we theoretically show how the regions of attraction influence the stability of the system subject to external perturbations. Generalizing the bistable system into coupled networks, we also find the role of network parameters in shaping the regions of attraction. Our findings shed light on the intricate interplay between brain networks and epileptic activity, offering mechanistic insights into potential avenues for more predictable treatments.

eess.SY↗

arXiv2024

Universal Differential Equations as a Common Modeling Language for Neuroscience

The unprecedented availability of large-scale datasets in neuroscience has spurred the exploration of artificial deep neural networks (DNNs) both as empirical tools and as models of natural neural systems. Their appeal lies in their ability to approximate arbitrary functions directly from observations, circumventing the need for cumbersome mechanistic modeling. However, without appropriate constraints, DNNs risk producing implausible models, diminishing their scientific value. Moreover, the interpretability of DNNs poses a significant challenge, particularly with the adoption of more complex expressive architectures. In this perspective, we argue for universal differential equations (UDEs) as a unifying approach for model development and validation in neuroscience. UDEs view differential equations as parameterizable, differentiable mathematical objects that can be augmented and trained with scalable deep learning techniques. This synergy facilitates the integration of decades of extensive literature in calculus, numerical analysis, and neural modeling with emerging advancements in AI into a potent framework. We provide a primer on this burgeoning topic in scientific machine learning and demonstrate how UDEs fill in a critical gap between mechanistic, phenomenological, and data-driven models in neuroscience. We outline a flexible recipe for modeling neural systems with UDEs and discuss how they can offer principled solutions to inherent challenges across diverse neuroscience applications such as understanding neural computation, controlling neural systems, neural decoding, and normative modeling.

cs.CE↗

arXiv2024

Efficient Deep Reinforcement Learning with Predictive Processing Proximal Policy Optimization

Advances in reinforcement learning (RL) often rely on massive compute resources and remain notoriously sample inefficient. In contrast, the human brain is able to efficiently learn effective control strategies using limited resources. This raises the question whether insights from neuroscience can be used to improve current RL methods. Predictive processing is a popular theoretical framework which maintains that the human brain is actively seeking to minimize surprise. We show that recurrent neural networks which predict their own sensory states can be leveraged to minimise surprise, yielding substantial gains in cumulative reward. Specifically, we present the Predictive Processing Proximal Policy Optimization (P4O) agent; an actor-critic reinforcement learning agent that applies predictive processing to a recurrent variant of the PPO algorithm by integrating a world model in its hidden state. Even without hyperparameter tuning, P4O significantly outperforms a baseline recurrent variant of the PPO algorithm on multiple Atari games using a single GPU. It also outperforms other state-of-the-art agents given the same wall-clock time and exceeds human gamer performance on multiple games including Seaquest, which is a particularly challenging environment in the Atari domain. Altogether, our work underscores how insights from the field of neuroscience may support the development of more capable and efficient artificial agents.

cs.LG↗

arXiv2023

Constrained Parameter Inference as a Principle for Learning

Learning in neural networks is often framed as a problem in which targeted error signals are directly propagated to parameters and used to produce updates that induce more optimal network behaviour. Backpropagation of error (BP) is an example of such an approach and has proven to be a highly successful application of stochastic gradient descent to deep neural networks. We propose constrained parameter inference (COPI) as a new principle for learning. The COPI approach assumes that learning can be set up in a manner where parameters infer their own values based upon observations of their local neuron activities. We find that this estimation of network parameters is possible under the constraints of decorrelated neural inputs and top-down perturbations of neural states for credit assignment. We show that the decorrelation required for COPI allows learning at extremely high learning rates, competitive with that of adaptive optimizers, as used by BP. We further demonstrate that COPI affords a new approach to feature analysis and network compression. Finally, we argue that COPI may shed new light on learning in biological networks given the evidence for decorrelation in the brain.

cs.NE↗

arXiv2023

Gradient-adjusted Incremental Target Propagation Provides Effective Credit Assignment in Deep Neural Networks

Many of the recent advances in the field of artificial intelligence have been fueled by the highly successful backpropagation of error (BP) algorithm, which efficiently solves the credit assignment problem in artificial neural networks. However, it is unlikely that BP is implemented in its usual form within biological neural networks, because of its reliance on non-local information in propagating error gradients. Since biological neural networks are capable of highly efficient learning and responses from BP trained models can be related to neural responses, it seems reasonable that a biologically viable approximation of BP underlies synaptic plasticity in the brain. Gradient-adjusted incremental target propagation (GAIT-prop or GP for short) has recently been derived directly from BP and has been shown to successfully train networks in a more biologically plausible manner. However, so far, GP has only been shown to work on relatively low-dimensional problems, such as handwritten-digit recognition. This work addresses some of the scaling issues in GP and shows it to perform effective multi-layer credit assignment in deeper networks and on the much more challenging ImageNet dataset.

cs.LG↗

arXiv2022

Learning Policies for Continuous Control via Transition Models

It is doubtful that animals have perfect inverse models of their limbs (e.g., what muscle contraction must be applied to every joint to reach a particular location in space). However, in robot control, moving an arm's end-effector to a target position or along a target trajectory requires accurate forward and inverse models. Here we show that by learning the transition (forward) model from interaction, we can use it to drive the learning of an amortized policy. Hence, we revisit policy optimization in relation to the deep active inference framework and describe a modular neural network architecture that simultaneously learns the system dynamics from prediction errors and the stochastic policy that generates suitable continuous control commands to reach a desired reference position. We evaluated the model by comparing it against the baseline of a linear quadratic regulator, and conclude with additional steps to take toward human-like motor control.

cs.RO↗

arXiv2022

How does artificial intelligence contribute to iEEG research?

Artificial intelligence (AI) is a fast-growing field focused on modeling and machine implementation of various cognitive functions with an increasing number of applications in computer vision, text processing, robotics, neurotechnology, bio-inspired computing and others. In this chapter, we describe how AI methods can be applied in the context of intracranial electroencephalography (iEEG) research. IEEG data is unique as it provides extremely high-quality signals recorded directly from brain tissue. Applying advanced AI models to these data carries the potential to further our understanding of many fundamental questions in neuroscience. At the same time, as an invasive technique, iEEG lends itself well to long-term, mobile brain-computer interface applications, particularly for communication in severely paralyzed individuals. We provide a detailed overview of these two research directions in the application of AI techniques to iEEG. That is, (1) the development of computational models that target fundamental questions about the neurobiological nature of cognition (AI-iEEG for neuroscience) and (2) applied research on monitoring and identification of event-driven brain states for the development of clinical brain-computer interface systems (AI-iEEG for neurotechnology). We explain key machine learning concepts, specifics of processing and modeling iEEG data and details of state-of-the-art iEEG-based neurotechnology and brain-computer interfaces.

q-bio.NC↗

arXiv2021

Explainable Deep Learning: A Field Guide for the Uninitiated

Deep neural networks (DNNs) have become a proven and indispensable machine learning tool. As a black-box model, it remains difficult to diagnose what aspects of the model's input drive the decisions of a DNN. In countless real-world domains, from legislation and law enforcement to healthcare, such diagnosis is essential to ensure that DNN decisions are driven by aspects appropriate in the context of its use. The development of methods and studies enabling the explanation of a DNN's decisions has thus blossomed into an active, broad area of research. A practitioner wanting to study explainable deep learning may be intimidated by the plethora of orthogonal directions the field has taken. This complexity is further exacerbated by competing definitions of what it means ``to explain'' the actions of a DNN and to evaluate an approach's ``ability to explain''. This article offers a field guide to explore the space of explainable deep learning aimed at those uninitiated in the field. The field guide: i) Introduces three simple dimensions defining the space of foundational methods that contribute to explainable deep learning, ii) discusses the evaluations for model explanations, iii) places explainability in the context of other related deep learning research areas, and iv) finally elaborates on user-oriented explanation designing and potential future directions on explainable deep learning. We hope the guide is used as an easy-to-digest starting point for those just embarking on research in this field.

cs.LG↗

arXiv2021

Neuroscience-inspired perception-action in robotics: applying active inference for state estimation, control and self-perception

Unlike robots, humans learn, adapt and perceive their bodies by interacting with the world. Discovering how the brain represents the body and generates actions is of major importance for robotics and artificial intelligence. Here we discuss how neuroscience findings open up opportunities to improve current estimation and control algorithms in robotics. In particular, how active inference, a mathematical formulation of how the brain resists a natural tendency to disorder, provides a unified recipe to potentially solve some of the major challenges in robotics, such as adaptation, robustness, flexibility, generalization and safe interaction. This paper summarizes some experiments and lessons learned from developing such a computational model on real embodied platforms, i.e., humanoid and industrial robots. Finally, we showcase the limitations and challenges that we are still facing to give robots human-like perception

cs.RO↗

arXiv2021

Automatic structured variational inference

Stochastic variational inference offers an attractive option as a default method for differentiable probabilistic programming. However, the performance of the variational approach depends on the choice of an appropriate variational family. Here, we introduce automatic structured variational inference (ASVI), a fully automated method for constructing structured variational families, inspired by the closed-form update in conjugate Bayesian models. These convex-update families incorporate the forward pass of the input probabilistic program and can therefore capture complex statistical dependencies. Convex-update families have the same space and time complexity as the input probabilistic program and are therefore tractable for a very large family of models including both continuous and discrete variables. We validate our automatic variational method on a wide range of low- and high-dimensional inference problems. We find that ASVI provides a clear improvement in performance when compared with other popular approaches such as the mean-field approach and inverse autoregressive flows. We provide an open source implementation of ASVI in TensorFlow Probability.

stat.ML↗