arXiv subjects

Omri Barak

Publications and source records attributed to Omri Barak.

18 recordsLinked to original sources

arXiv2026

Learning reveals invisible structure in low-rank RNNs

Learning in neural systems arises from synaptic changes that reshape the representations underlying behavior. While low-rank recurrent neural networks (RNNs) have emerged as a powerful framework for linking connectivity to function, a theoretical understanding of their learning process remains elusive. Here, we extend the low-rank framework from activity to learning by deriving gradient-descent dynamics directly in a reduced overlap space. We formulate a closed-form, low-dimensional system of ODEs that governs learning in this space, exact for linear RNNs and asymptotically exact for nonlinear RNNs in the large-$N$ Gaussian limit. Central to our analysis is a distinction between two classes of overlaps: loss-visible overlaps, which fully determine network activity, output, and loss, and loss-invisible overlaps, which do not affect function but are required to describe learning. We illustrate the consequences of this decomposition through two phenomena. First, we show that learning can serve as a perturbation that exposes differences in connectivity between functionally equivalent networks. Second, we show that loss-invisible overlaps can act as memory variables that encode training history, and characterize the conditions under which this occurs. Finally, we present several testable predictions for biological learning experiments derived from our theory.

cs.LG↗

arXiv2025

Prediction horizon shapes representations in predictive learning

Predictive learning has emerged as a central paradigm for training models across diverse data domains and is increasingly viewed as a foundation for modern artificial intelligence. A common intuition for this success is that accurate prediction requires models to capture the underlying dynamics of the environment, leading to the emergence of structured world models. However, predictive learning does not universally yield such representations, and a mechanistic account of when and why it does remains incomplete. In this work, we identify the prediction horizon as a critical, but often implicit, component of predictive learning objectives. We show that increasing the prediction horizon fundamentally shapes the effective structure of the learning problem. In a minimal setting, we demonstrate both theoretically and empirically that the model's implicit biases interact with this structural change to recover the latent geometry of the task. We then extend these empirical results to nonlinear architectures and more complex datasets, where similar phenomena persist. These findings provide a principled explanation for the emergence of structured representations in predictive learning paradigms and clarify the conditions under which such representations should be expected.

cs.LG↗

arXiv2025

Finding separatrices of dynamical flows with Deep Koopman Eigenfunctions

Many natural systems, including neural circuits involved in decision making, are modeled as high-dimensional dynamical systems with multiple stable states. While existing analytical tools primarily describe behavior near stable equilibria, characterizing separatrices--the manifolds that delineate boundaries between different basins of attraction--remains challenging, particularly in high-dimensional settings. Here, we introduce a numerical framework leveraging Koopman Theory combined with Deep Neural Networks to effectively characterize separatrices. Specifically, we approximate Koopman Eigenfunctions (KEFs) associated with real positive eigenvalues, which vanish precisely at the separatrices. Utilizing these scalar KEFs, optimization methods efficiently locate separatrices even in complex systems. We demonstrate our approach on synthetic benchmarks, ecological network models, and high-dimensional recurrent neural networks trained on either neuroscience-inspired tasks or fit to real neural data. Moreover, we illustrate the practical utility of our method by designing optimal perturbations that can shift systems across separatrices, enabling predictions relevant to optogenetic stimulation experiments in neuroscience.

cs.LG↗

arXiv2025

Learning Dynamics of RNNs in Closed-Loop Environments

Recurrent neural networks (RNNs) trained on neuroscience-inspired tasks offer powerful models of brain computation. However, typical training paradigms rely on open-loop, supervised settings, whereas real-world learning unfolds in closed-loop environments. Here, we develop a mathematical theory describing the learning dynamics of linear RNNs trained in closed-loop contexts. We first demonstrate that two otherwise identical RNNs, trained in either closed- or open-loop modes, follow markedly different learning trajectories. To probe this divergence, we analytically characterize the closed-loop case, revealing distinct stages aligned with the evolution of the training loss. Specifically, we show that the learning dynamics of closed-loop RNNs, in contrast to open-loop ones, are governed by an interplay between two competing objectives: short-term policy improvement and long-term stability of the agent-environment interaction. Finally, we apply our framework to a realistic motor control task, highlighting its broader applicability. Taken together, our results underscore the importance of modeling closed-loop dynamics in a biologically plausible setting.

cs.LG↗

arXiv2024

Aligned and oblique dynamics in recurrent neural networks

The relation between neural activity and behaviorally relevant variables is at the heart of neuroscience research. When strong, this relation is termed a neural representation. There is increasing evidence, however, for partial dissociations between activity in an area and relevant external variables. While many explanations have been proposed, a theoretical framework for the relationship between external and internal variables is lacking. Here, we utilize recurrent neural networks (RNNs) to explore the question of when and how neural dynamics and the network's output are related from a geometrical point of view. We find that training RNNs can lead to two dynamical regimes: dynamics can either be aligned with the directions that generate output variables, or oblique to them. We show that the choice of readout weight magnitude before training can serve as a control knob between the regimes, similar to recent findings in feedforward networks. These regimes are functionally distinct. Oblique networks are more heterogeneous and suppress noise in their output directions. They are furthermore more robust to perturbations along the output directions. Crucially, the oblique regime is specific to recurrent (but not feedforward) networks, arising from dynamical stability considerations. Finally, we show that tendencies towards the aligned or the oblique regime can be dissociated in neural recordings. Altogether, our results open a new perspective for interpreting neural activity by relating network dynamics and their output.

q-bio.NC↗

arXiv2024

When predict can also explain: few-shot prediction to select better neural latents

Latent variable models serve as powerful tools to infer underlying dynamics from observed neural activity. Ideally, the inferred dynamics should align with true ones. However, due to the absence of ground truth data, prediction benchmarks are often employed as proxies. One widely-used method, $\textit{co-smoothing}$, involves jointly estimating latent variables and predicting observations along held-out channels to assess model performance. In this study, we reveal the limitations of the co-smoothing prediction framework and propose a remedy. Using a student-teacher setup, we demonstrate that models with high co-smoothing can have arbitrary extraneous dynamics in their latent representations. To address this, we introduce a secondary metric -- $\textit{few-shot co-smoothing}$, performing regression from the latent variables to held-out neurons in the data using fewer trials. Our results indicate that among models with near-optimal co-smoothing, those with extraneous dynamics underperform in the few-shot co-smoothing compared to `minimal' models that are devoid of such dynamics. We provide analytical insights into the origin of this phenomenon and further validate our findings on four standard neural datasets using a state-of-the-art method: STNDT. In the absence of ground truth, we suggest a novel measure to validate our approach. By cross-decoding the latent variables of all model pairs with high co-smoothing, we identify models with minimal extraneous dynamics. We find a correlation between few-shot co-smoothing performance and this new measure. In summary, we present a novel prediction metric designed to yield latent variables that more accurately reflect the ground truth, offering a significant improvement for latent dynamics inference.

cs.LG↗

arXiv2021

Charting and navigating the space of solutions for recurrent neural networks

Recurrent Neural Networks (RNNs) were recently successfully used to model the way neural activity drives task-related behavior in animals, operating under the implicit assumption that the obtained solutions are universal. Observations in both neuroscience and machine learning challenge this assumption. Animals can approach a given task with a variety of strategies, and training machine learning algorithms introduces the phenomenon of underspecification. These observations imply that every task is associated with a space of solutions. To date, the structure of this space is not understood, limiting the approach of comparing RNNs with neural data. Here, we characterize the space of solutions associated with various tasks. We first study a simple two-neuron network on a task that leads to multiple solutions. We trace the nature of the final solution back to the network's initial connectivity and identify discrete dynamical regimes that underlie this diversity. We then examine three neuroscience-inspired tasks: Delayed and interval discrimination, and Time reproduction. For each task, we find a rich set of solutions. Variability can be found directly in the neural activity of the networks, and additionally by testing the trained networks' ability to extrapolate, as a perturbation to a system often reveals hidden structure. Furthermore, we relate extrapolation patterns to specific dynamical objects and effective algorithms found by the networks. We introduce a tool to derive the reduced dynamics of networks by generating a compact directed graph describing the essence of the dynamics with regards to behavioral inputs and outputs. Using this representation, we can partition the solutions to each task into a handful of types and partially predict them from neural features. Our results shed light on the concept of the space of solutions and its uses in Machine learning and in Neuroscience.

q-bio.NC↗

arXiv2021

Cancer Progression as a Learning Process

Drug resistance and metastasis - the major complications in cancer - both entail adaptation of cancer cells to stress, whether a drug or a lethal new environment. Intriguingly, these adaptive processes share similar features that cannot be explained by a pure Darwinian scheme, including dormancy, increased heterogeneity, and stress-induced plasticity. Here, we propose that learning theory offers a framework to explain these features and may shed light on these two intricate processes. In this framework, learning is performed at the single cell level, by stress-driven exploratory trial-and-error. Such a process is not contingent on pre-existing pathways but on a random search for a state that diminishes the stress. We review underlying mechanisms that may support this search, and show by using a learning model that such exploratory adaptation is feasible in a high dimensional system as the cell. At the population level, we view the tissue as a network of exploring agents that communicate and restrain cancer formation in health. In this view, disease results from the breakdown of homeostasis between cellular exploratory drive and tissue homeostasis.

q-bio.MN↗

arXiv2021

The interplay between randomness and structure during learning in RNNs

Recurrent neural networks (RNNs) trained on low-dimensional tasks have been widely used to model functional biological networks. However, the solutions found by learning and the effect of initial connectivity are not well understood. Here, we examine RNNs trained using gradient descent on different tasks inspired by the neuroscience literature. We find that the changes in recurrent connectivity can be described by low-rank matrices, despite the unconstrained nature of the learning algorithm. To identify the origin of the low-rank structure, we turn to an analytically tractable setting: training a linear RNN on a simplified task. We show how the low-dimensional task structure leads to low-rank changes to connectivity. This low-rank structure allows us to explain and quantify the phenomenon of accelerated learning in the presence of random initial connectivity. Altogether, our study opens a new perspective to understanding trained RNNs in terms of both the learning process and the resulting network structure.

q-bio.NC↗

arXiv2020

Quality of internal representation shapes learning performance in feedback neural networks

A fundamental feature of complex biological systems is the ability to form feedback interactions with their environment. A prominent model for studying such interactions is reservoir computing, where learning acts on low-dimensional bottlenecks. Despite the simplicity of this learning scheme, the factors contributing to or hindering the success of training in reservoir networks are in general not well understood. In this work, we study non-linear feedback networks trained to generate a sinusoidal signal, and analyze how learning performance is shaped by the interplay between internal network dynamics and target properties. By performing exact mathematical analysis of linearized networks, we predict that learning performance is maximized when the target is characterized by an optimal, intermediate frequency which monotonically decreases with the strength of the internal reservoir connectivity. At the optimal frequency, the reservoir representation of the target signal is high-dimensional, de-synchronized, and thus maximally robust to noise. We show that our predictions successfully capture the qualitative behaviour of performance in non-linear networks. Moreover, we find that the relationship between internal representations and performance can be further exploited in trained non-linear networks to explain behaviours which do not have a linear counterpart. Our results indicate that a major determinant of learning success is the quality of the internal representation of the target, which in turn is shaped by an interplay between parameters controlling the internal network and those defining the task.

q-bio.NC↗

arXiv2020

Mapping low-dimensional dynamics to high-dimensional neural activity: A derivation of the ring model from the neural engineering framework

Empirical estimates of the dimensionality of neural population activity are often much lower than the population size. Similar phenomena are also observed in trained and designed neural network models. These experimental and computational results suggest that mapping low-dimensional dynamics to high-dimensional neural space is a common feature of cortical computation. Despite the ubiquity of this observation, the constraints arising from such mapping are poorly understood. Here we consider a specific example of mapping low-dimensional dynamics to high-dimensional neural activity -- the neural engineering framework. We analytically solve the framework for the classic ring model -- a neural network encoding a static or dynamic angular variable. Our results provide a complete characterization of the success and failure modes for this model. Based on similarities between this and other frameworks, we speculate that these results could apply to more general scenarios.

q-bio.NC↗

arXiv2020

Implementing Inductive bias for different navigation tasks through diverse RNN attractors

Navigation is crucial for animal behavior and is assumed to require an internal representation of the external environment, termed a cognitive map. The precise form of this representation is often considered to be a metric representation of space. An internal representation, however, is judged by its contribution to performance on a given task, and may thus vary between different types of navigation tasks. Here we train a recurrent neural network that controls an agent performing several navigation tasks in a simple environment. To focus on internal representations, we split learning into a task-agnostic pre-training stage that modifies internal connectivity and a task-specific Q learning stage that controls the network's output. We show that pre-training shapes the attractor landscape of the networks, leading to either a continuous attractor, discrete attractors or a disordered state. These structures induce bias onto the Q-Learning phase, leading to a performance pattern across the tasks corresponding to metric and topological regularities. By combining two types of networks in a modular structure, we could get better performance for both regularities. Our results show that, in recurrent networks, inductive bias takes the form of attractor landscapes -- which can be shaped by pre-training and analyzed using dynamical systems methods. Furthermore, we demonstrate that non-metric representations are useful for navigation tasks, and their combination with metric representation leads to flexibile multiple-task learning.

q-bio.NC↗

arXiv2019

Dynamics of random recurrent networks with correlated low-rank structure

A given neural network in the brain is involved in many different tasks. This implies that, when considering a specific task, the network's connectivity contains a component which is related to the task and another component which can be considered random. Understanding the interplay between the structured and random components, and their effect on network dynamics and functionality is an important open question. Recent studies addressed the co-existence of random and structured connectivity, but considered the two parts to be uncorrelated. This constraint limits the dynamics and leaves the random connectivity non-functional. Algorithms that train networks to perform specific tasks typically generate correlations between structure and random connectivity. Here we study nonlinear networks with correlated structured and random components, assuming the structure to have a low rank. We develop an analytic framework to establish the precise effect of the correlations on the eigenvalue spectrum of the joint connectivity. We find that the spectrum consists of a bulk and multiple outliers, whose location is predicted by our theory. Using mean-field theory, we show that these outliers directly determine both the fixed points of the system and their stability. Taken together, our analysis elucidates how correlations allow structured and random connectivity to synergistically extend the range of computations available to networks.

q-bio.NC↗

arXiv2019

Understanding and Controlling Memory in Recurrent Neural Networks

To be effective in sequential data processing, Recurrent Neural Networks (RNNs) are required to keep track of past events by creating memories. While the relation between memories and the network's hidden state dynamics was established over the last decade, previous works in this direction were of a predominantly descriptive nature focusing mainly on locating the dynamical objects of interest. In particular, it remained unclear how dynamical observables affect the performance, how they form and whether they can be manipulated. Here, we utilize different training protocols, datasets and architectures to obtain a range of networks solving a delayed classification task with similar performance, alongside substantial differences in their ability to extrapolate for longer delays. We analyze the dynamics of the network's hidden state, and uncover the reasons for this difference. Each memory is found to be associated with a nearly steady state of the dynamics which we refer to as a 'slow point'. Slow point speeds predict extrapolation performance across all datasets, protocols and architectures tested. Furthermore, by tracking the formation of the slow points we are able to understand the origin of differences between training protocols. Finally, we propose a novel regularization technique that is based on the relation between hidden state speeds and memory longevity. Our technique manipulates these speeds, thereby leading to a dramatic improvement in memory robustness over time, and could pave the way for a new class of regularization methods.

cs.LG↗

arXiv2019

Repeated sequential learning increases memory capacity via effective decorrelation in a recurrent neural network

Memories in neural system are shaped through the interplay of neural and learning dynamics under external inputs. By introducing a simple local learning rule to a neural network, we found that the memory capacity is drastically increased by sequentially repeating the learning steps of input-output mappings. The origin of this enhancement is attributed to the generation of a Psuedo-inverse correlation in the connectivity. This is associated with the emergence of spontaneous activity that intermittently exhibits neural patterns corresponding to embedded memories. Stablization of memories is achieved by a distinct bifurcation from the spontaneous activity under the application of each input.

nlin.AO↗

arXiv2019

One step back, two steps forward: interference and learning in recurrent neural networks

Artificial neural networks, trained to perform cognitive tasks, have recently been used as models for neural recordings from animals performing these tasks. While some progress has been made in performing such comparisons, the evolution of network dynamics throughout learning remains unexplored. This is paralleled by an experimental focus on recording from trained animals, with few studies following neural activity throughout training. In this work, we address this gap in the realm of artificial networks by analyzing networks that are trained to perform memory and pattern generation tasks. The functional aspect of these tasks corresponds to dynamical objects in the fully trained network - a line attractor or a set of limit cycles for the two respective tasks. We use these dynamical objects as anchors to study the effect of learning on their emergence. We find that the sequential nature of learning has major consequences for the learning trajectory and its final outcome. Specifically, we show that Least Mean Squares (LMS), a simple gradient descent suggested as a biologically plausible version of the FORCE algorithm, is constantly obstructed by forgetting, which is manifested as the destruction of dynamical objects from previous trials. The degree of interference is determined by the correlation between different trials. We show which specific ingredients of FORCE avoid this phenomenon. Overall, this difference results in convergence that is orders of magnitude slower for LMS. Learning implies accumulating information across multiple trials to form the overall concept of the task. Our results show that interference between trials can greatly affect learning, in a learning rule dependent manner. These insights can help design experimental protocols that minimize such interference, and possibly infer underlying learning rules by observing behavior and neural activity throughout learning.

q-bio.NC↗

arXiv2019

Stable memory with unstable synapses

What is the physiological basis of long-term memory? The prevailing view in neuroscience attributes changes in synaptic efficacy to memory acquisition. This view implies that stable memories correspond to stable connectivity patterns. However, an increasing body of experimental evidence points to significant, activity-independent dynamics in synaptic strengths. Motivated by these observations, we explore the possibility of memory storage within a global component of network connectivity, while individual connections fluctuate. We find a simple and general principle, stemming from stability arguments, that links eigenvalues in the complex plane to memories. Specifically, imaginary-coded memories are more resilient to noise and homeostatic plasticity than their real-coded counterparts. Memory representations are stored as time-varying attractors in neural state-space and support associative retrieval of learned information. Our results suggest a link between the properties of learning rules and those of network-level memory representations, and point at measurable signatures to be sought in experimental data.

q-bio.NC↗

arXiv2016

Local Dynamics in Trained Recurrent Neural Networks

Learning a task induces connectivity changes in neural circuits, thereby changing their dynamics. To elucidate task related neural dynamics we study trained Recurrent Neural Networks. We develop a Mean Field Theory for Reservoir Computing networks trained to have multiple fixed point attractors. Our main result is that the dynamics of the network's output in the vicinity of attractors is governed by a low order linear Ordinary Differential Equation. Stability of the resulting ODE can be assessed, predicting training success or failure. As a consequence, networks of Rectified Linear (RLU) and of sigmoidal nonlinearities are shown to have diametrically different properties when it comes to learning attractors. Furthermore, a characteristic time constant, which remains finite at the edge of chaos, offers an explanation of the network's output robustness in the presence of variability of the internal neural dynamics. Finally, the proposed theory predicts state dependent frequency selectivity in network response.

q-bio.NC↗