arXiv subjects

Paolo Muratore

Publications and source records attributed to Paolo Muratore.

9 recordsLinked to original sources

arXiv2026

Extracting Governing Equations from Latent Dynamics via Multi-View Contrastive Learning

Identifying latent dynamical systems from noisy, high-dimensional measurements is a central problem at the intersection of representation learning, system identification, and scientific discovery. We present DYSCO, a multi-view temporal contrastive learning algorithm that jointly recovers latent trajectories and the governing dynamics from such observations, by leveraging multiple independent noisy views of the same underlying process to disentangle signal from noise. By parameterizing the dynamics in a structured functional basis, our framework further enables symbolic recovery of the governing equations within an affine gauge. We offer theoretical guarantees for strong identification up to an affine indeterminacy, extending prior identifiability results to the realistic setting of noisy nonlinear observations. Empirically, we demonstrate accurate recovery of both latent trajectories and flow fields across a diverse set of dynamical regimes (e.g., chaotic, oscillatory, and metastable) under both Gaussian and Poisson observation noise, the latter being particularly relevant for neural recordings.

cs.LG↗

arXiv2026

Stretching Beyond the Obvious: A Gradient-Free Framework to Unveil the Hidden Landscape of Visual Invariance

Uncovering which feature combinations are encoded by visual units is critical to understanding how images are transformed into representations that support recognition. While existing feature visualization approaches typically infer a unit's most exciting images, this is insufficient to reveal the manifold of transformations under which responses remain invariant, which is critical to generalization in vision. Here we introduce Stretch-and-Squeeze (SnS), a model-agnostic, gradient-free framework to systematically characterize a unit's maximally invariant stimuli, and its vulnerability to adversarial perturbations, in both biological and artificial visual systems. SnS frames these transformations as bi-objective optimization problems. To probe invariance, SnS seeks image perturbations that maximally alter (stretch) the representation of a reference stimulus in a given processing stage while preserving unit activation downstream (squeeze). To probe adversarial sensitivity, stretching and squeezing are reversed to maximally perturb unit activation while minimizing changes to the upstream representation. Applied to CNNs, SnS revealed invariant transformations that were farther from a reference image in pixel-space than those produced by affine transformations, while more strongly preserving the target unit's response. The discovered invariant images differed depending on the stage of the image representation used for optimization: pixel-level changes primarily affected luminance and contrast, while stretching mid- and late-layer representations mainly altered texture and pose. By measuring how well the hierarchical invariant images obtained for L2 robust networks were classified by humans and other observer networks, we discovered a substantial drop in their interpretability when the representation was stretched in deep layers, while the opposite trend was found for standard models.

cs.CV↗

arXiv2024

Learning fast changing slow in spiking neural networks

Reinforcement learning (RL) faces substantial challenges when applied to real-life problems, primarily stemming from the scarcity of available data due to limited interactions with the environment. This limitation is exacerbated by the fact that RL often demands a considerable volume of data for effective learning. The complexity escalates further when implementing RL in recurrent spiking networks, where inherent noise introduced by spikes adds a layer of difficulty. Life-long learning machines must inherently resolve the plasticity-stability paradox. Striking a balance between acquiring new knowledge and maintaining stability is crucial for artificial agents. To address this challenge, we draw inspiration from machine learning technology and introduce a biologically plausible implementation of proximal policy optimization, referred to as lf-cs (learning fast changing slow). Our approach results in two notable advancements: firstly, the capacity to assimilate new information into a new policy without requiring alterations to the current policy; and secondly, the capability to replay experiences without experiencing policy divergence. Furthermore, when contrasted with other experience replay (ER) techniques, our method demonstrates the added advantage of being computationally efficient in an online setting. We demonstrate that the proposed methodology enhances the efficiency of learning, showcasing its potential impact on neuromorphic and real-world applications.

cs.NE↗

arXiv2022

Beyond spiking networks: the computational advantages of dendritic amplification and input segregation

The brain can efficiently learn a wide range of tasks, motivating the search for biologically inspired learning rules for improving current artificial intelligence technology. Most biological models are composed of point neurons, and cannot achieve the state-of-the-art performances in machine learning. Recent works have proposed that segregation of dendritic input (neurons receive sensory information and higher-order feedback in segregated compartments) and generation of high-frequency bursts of spikes would support error backpropagation in biological neurons. However, these approaches require propagating errors with a fine spatio-temporal structure to the neurons, which is unlikely to be feasible in a biological network. To relax this assumption, we suggest that bursts and dendritic input segregation provide a natural support for biologically plausible target-based learning, which does not require error propagation. We propose a pyramidal neuron model composed of three separated compartments. A coincidence mechanism between the basal and the apical compartments allows for generating high-frequency bursts of spikes. This architecture allows for a burst-dependent learning rule, based on the comparison between the target bursting activity triggered by the teaching signal and the one caused by the recurrent connections, providing the support for target-based learning. We show that this framework can be used to efficiently solve spatio-temporal tasks, such as the store and recall of 3D trajectories. Finally, we suggest that this neuronal architecture naturally allows for orchestrating ``hierarchical imitation learning'', enabling the decomposition of challenging long-horizon decision-making tasks into simpler subtasks. This can be implemented in a two-level network, where the high-network acts as a ``manager'' and produces the contextual signal for the low-network, the ``worker''.

q-bio.NC↗

arXiv2022

Prune and distill: similar reformatting of image information along rat visual cortex and deep neural networks

Visual object recognition has been extensively studied in both neuroscience and computer vision. Recently, the most popular class of artificial systems for this task, deep convolutional neural networks (CNNs), has been shown to provide excellent models for its functional analogue in the brain, the ventral stream in visual cortex. This has prompted questions on what, if any, are the common principles underlying the reformatting of visual information as it flows through a CNN or the ventral stream. Here we consider some prominent statistical patterns that are known to exist in the internal representations of either CNNs or the visual cortex and look for them in the other system. We show that intrinsic dimensionality (ID) of object representations along the rat homologue of the ventral stream presents two distinct expansion-contraction phases, as previously shown for CNNs. Conversely, in CNNs, we show that training results in both distillation and active pruning (mirroring the increase in ID) of low- to middle-level image information in single units, as representations gain the ability to support invariant discrimination, in agreement with previous observations in rat visual cortex. Taken together, our findings suggest that CNNs and visual cortex share a similarly tight relationship between dimensionality expansion/reduction of object representations and reformatting of image information.

q-bio.NC↗

arXiv2022

Burst-dependent plasticity and dendritic amplification support target-based learning and hierarchical imitation learning

The brain can learn to solve a wide range of tasks with high temporal and energetic efficiency. However, most biological models are composed of simple single compartment neurons and cannot achieve the state-of-art performances of artificial intelligence. We propose a multi-compartment model of pyramidal neuron, in which bursts and dendritic input segregation give the possibility to plausibly support a biological target-based learning. In target-based learning, the internal solution of a problem (a spatio temporal pattern of bursts in our case) is suggested to the network, bypassing the problems of error backpropagation and credit assignment. Finally, we show that this neuronal architecture naturally support the orchestration of hierarchical imitation learning, enabling the decomposition of challenging long-horizon decision-making tasks into simpler subtasks.

q-bio.NC↗

arXiv2021

Error-based or target-based? A unifying framework for learning in recurrent spiking networks

Learning in biological or artificial networks means changing the laws governing the network dynamics in order to better behave in a specific situation. In the field of supervised learning, two complementary approaches stand out: error-based and target-based learning. However, there exists no consensus on which is better suited for which task, and what is the most biologically plausible. Here we propose a comprehensive theoretical framework that includes these two frameworks as special cases. This novel theoretical formulation offers major insights into the differences between the two approaches. In particular, we show how target-based naturally emerges from error-based when the number of constraints on the target dynamics, and as a consequence on the internal network dynamics, is comparable to the degrees of freedom of the network. Moreover, given the experimental evidences on the relevance that spikes have in biological networks, we investigate the role of coding with specific patterns of spikes by introducing a parameter that defines the tolerance to precise spike timing during learning. Our approach naturally lends itself to Imitation Learning (and Behavioral Cloning in particular) and we apply it to solve relevant closed-loop tasks such as the button-and-food task, and the 2D Bipedal Walker. We show that a high dimensionality feedback structure is extremely important when it is necessary to solve a task that requires retaining memory for a long time (button-and-food). On the other hand, we find that coding with specific patterns of spikes enables optimal performances in a motor task (the 2D Bipedal Walker). Finally, we show that our theoretical formulation suggests protocols to deduce the structure of learning feedback in biological networks.

q-bio.NC↗

arXiv2021

Target spiking patterns enable efficient and biologically plausible learning for complex temporal tasks

Recurrent spiking neural networks (RSNN) in the human brain learn to perform a wide range of perceptual, cognitive and motor tasks very efficiently in terms of energy consumption and requires very few examples. This motivates the search for biologically inspired learning rules for RSNNs to improve our understanding of brain computation and the efficiency of artificial intelligence. Several spiking models and learning rules have been proposed, but it remains a challenge to design RSNNs whose learning relies on biologically plausible mechanisms and are capable of solving complex temporal tasks. In this paper, we derive a learning rule, local to the synapse, from a simple mathematical principle, the maximization of the likelihood for the network to solve a specific task. We propose a novel target-based learning scheme in which the learning rule derived from likelihood maximization is used to mimic a specific spiking pattern that encodes the solution to complex temporal tasks. This method makes the learning extremely rapid and precise, outperforming state of the art algorithms for RSNNs. We demonstrate the capacity of our model to tackle several problems like learning multidimensional trajectories and solving the classical temporal XOR benchmark. Finally, we show that an online approximation of the gradient ascent, in addition to guaranteeing complete locality in time and space, allows learning after very few presentations of the target output. Our model can be applied to different types of biological neurons. The analytically derived plasticity learning rule is specific to each neuron model and can produce a theoretical prediction for experimental validation.

q-bio.NC↗

arXiv2020

Analysis and Model of Cortical Slow Waves Acquired with Optical Techniques

Slow waves (SWs) are spatio-temporal patterns of cortical activity that occur both during natural sleep and anesthesia and are preserved across species. Even though electrophysiological recordings have been largely used to characterize brain states, they are limited in the spatial resolution and cannot target specific neuronal population. Recently, large-scale optical imaging techniques coupled with functional indicators overcame these restrictions, and new pipelines of analysis and novel approaches of SWs modelling are needed to extract relevant features of the spatio-temporal dynamics of SWs from these highly spatially resolved data-sets. Here we combined wide-field fluorescence microscopy and a transgenic mouse model expressing a calcium indicator (GCaMP6f) in excitatory neurons to study SW propagation over the meso-scale under ketamine anesthesia. We developed a versatile analysis pipeline to identify and quantify the spatio-temporal propagation of the SWs. Moreover, we designed a computational simulator based on a simple theoretical model, which takes into account the statistics of neuronal activity, the response of fluorescence proteins and the slow waves dynamics. The simulator was capable of synthesizing artificial signals that could reliably reproduce several features of the SWs observed in vivo, thus enabling a calibration tool for the analysis pipeline. Comparison of experimental and simulated data shows the robustness of the analysis tools and its potential to uncover mechanistic insights of the Slow Wave Activity (SWA).

q-bio.NC↗