arXiv subjects

Jason Z. Kim

Publications and source records attributed to Jason Z. Kim.

27 records · Page 2Linked to original sources

arXiv2020

Teaching Recurrent Neural Networks to Modify Chaotic Memories by Example

The ability to store and manipulate information is a hallmark of computational systems. Whereas computers are carefully engineered to represent and perform mathematical operations on structured data, neurobiological systems perform analogous functions despite flexible organization and unstructured sensory input. Recent efforts have made progress in modeling the representation and recall of information in neural systems. However, precisely how neural systems learn to modify these representations remains far from understood. Here we demonstrate that a recurrent neural network (RNN) can learn to modify its representation of complex information using only examples, and we explain the associated learning mechanism with new theory. Specifically, we drive an RNN with examples of translated, linearly transformed, or pre-bifurcated time series from a chaotic Lorenz system, alongside an additional control signal that changes value for each example. By training the network to replicate the Lorenz inputs, it learns to autonomously evolve about a Lorenz-shaped manifold. Additionally, it learns to continuously interpolate and extrapolate the translation, transformation, and bifurcation of this representation far beyond the training data by changing the control signal. Finally, we provide a mechanism for how these computations are learned, and demonstrate that a single network can simultaneously learn multiple computations. Together, our results provide a simple but powerful mechanism by which an RNN can learn to manipulate internal representations of complex information, allowing for the principled study and precise design of RNNs.

cond-mat.dis-nn↗

arXiv2020

Models of communication and control for brain networks: distinctions, convergence, and future outlook

Recent advances in computational models of signal propagation and routing in the human brain have underscored the critical role of white matter structure. A complementary approach has utilized the framework of network control theory to better understand how white matter constrains the manner in which a region or set of regions can direct or control the activity of other regions. Despite the potential for both of these approaches to enhance our understanding of the role of network structure in brain function, little work has sought to understand the relations between them. Here, we seek to explicitly bridge computational models of communication and principles of network control in a conceptual review of the current literature. By drawing comparisons between communication and control models in terms of the level of abstraction, the dynamical complexity, the dependence on network attributes, and the interplay of multiple spatiotemporal scales, we highlight the convergence of and distinctions between the two frameworks. Based on the understanding of the intertwined nature of communication and control in human brain networks, this work provides an integrative perspective for the field and outlines exciting directions for future work.

q-bio.NC↗

arXiv2020

Supervised Chaotic Source Separation by a Tank of Water

Whether listening to overlapping conversations in a crowded room or recording the simultaneous electrical activity of millions of neurons, the natural world abounds with sparse measurements of complex overlapping signals that arise from dynamical processes. While tools that separate mixed signals into linear sources have proven necessary and useful, the underlying equational forms of most natural signals are unknown and nonlinear. Hence, there is a need for a framework that is general enough to extract sources without knowledge of their generating equations, and flexible enough to accommodate nonlinear, even chaotic, sources. Here we provide such a framework, where the sources are chaotic trajectories from independently evolving dynamical systems. We consider the mixture signal as the sum of two chaotic trajectories, and propose a supervised learning scheme that extracts the chaotic trajectories from their mixture. Specifically, we recruit a complex dynamical system as an intermediate processor that is constantly driven by the mixture. We then obtain the separated chaotic trajectories based on this intermediate system by training the proper output functions. To demonstrate the generalizability of this framework \emph{in silico}, we employ a tank of water as the intermediate system, and show its success in separating two-part mixtures of various chaotic trajectories. Finally, we relate the underlying mechanism of this method to the state-observer problem. This relation provides a quantitative theory that explains the performance of our method, such as why separation is difficult when two source signals are trajectories from the same chaotic system.

eess.SP↗

arXiv2019

Network structure of cascading neural systems predicts stimulus propagation and recovery

Many neural systems display cascading behavior characterized by uninterrupted sequences of neuronal firing. This gap precludes an understanding of how variations in network structure manifest in neural dynamics and either support or impinge upon information processing. Here, we develop a theoretical understanding of how network structure supports information processing through network dynamics, and we validate our theory with empirical data. Using a generalized spiking model and mathematical tools from linear systems theory, network control theory, and information theory, we show how network structure can be designed to temporally extend the propagation and recovery of certain stimulus patterns. Moreover, we observe cycles as structural and dynamic motifs that are prevalent in such networks. Broadly, our results demonstrate how cascading neural networks could contribute to cognitive faculties that require lasting activation of neuronal patterns, such as working memory or attention.

q-bio.NC↗

arXiv2019

Temporal sequences of brain activity at rest are constrained by white matter structure and modulated by cognitive demands

A diverse white matter network and finely tuned neuronal membrane properties allow the brain to transition seamlessly between cognitive states. However, it remains unclear how static structural connections guide the temporal progression of large-scale brain activity patterns in different cognitive states. Here, we analyze the brain's trajectories through a high-dimensional activity space at the level of single time point activity patterns from functional magnetic resonance imaging data acquired during passive visual fixation (rest) and an n-back working memory task. We find that specific state space trajectories, which represent temporal sequences of brain activity, are modulated by cognitive load and related to task performance. Using diffusion-weighted imaging acquired from the same subjects, we use tools from network control theory to show that linear spread of activity along white matter connections constrains the brain's state space trajectories at rest. Additionally, accounting for stimulus-driven visual inputs explains the different trajectories taken during the n-back task. We also used models of network rewiring to show that these findings are the result of non-trivial geometric and topological properties of white matter architecture. Finally, we examine associations between age and time-resolved brain state dynamics, revealing new insights into functional changes in the default mode and executive control networks. Overall, these results elucidate the structural underpinnings of cognitively and developmentally relevant spatiotemporal brain dynamics.

q-bio.NC↗

arXiv2019

A practical guide to methodological considerations in the controllability of structural brain networks

Predicting how the brain can be driven to specific states by means of internal or external control requires a fundamental understanding of the relationship between neural connectivity and activity. Network control theory is a powerful tool from the physical and engineering sciences that can provide insights regarding that relationship; it formalizes the study of how the dynamics of a complex system can arise from its underlying structure of interconnected units. Given the recent use of network control theory in neuroscience, it is now timely to offer a practical guide to methodological considerations in the controllability of structural brain networks. Here we provide a systematic overview of the framework, examine the impact of modeling choices on frequently studied control metrics, and suggest potentially useful theoretical extensions. We ground our discussions, numerical demonstrations, and theoretical advances in a dataset of high-resolution diffusion imaging with 730 diffusion directions acquired over approximately 1 hour of scanning from ten healthy young adults. Following a didactic introduction of the theory, we probe how a selection of modeling choices affects four common statistics: average controllability, modal controllability, minimum control energy, and optimal control energy. Next, we extend the current state of the art in two ways: first, by developing an alternative measure of structural connectivity that accounts for radial propagation of activity through abutting tissue, and second, by defining a complementary metric quantifying the complexity of the energy landscape of a system. We close with specific modeling recommendations and a discussion of methodological constraints.

q-bio.NC↗

arXiv2019

Design of Large Sequential Conformational Change in Mechanical Networks

From the complex motions of robots to the oxygen binding of hemoglobin, the function of many mechanical systems depends on large, coordinated movements of their components. Such movements arise from a network of physical interactions in the form of links that transmit forces between constituent elements. However, the principled design of specific movements is made difficult by the number and nonlinearity of interactions. Here, we model mechanical systems as linkages of rigid bonds (edges) connected by joints (nodes), and formulate a simple but powerful framework for designing full nonlinear coordinated motions using concepts from dynamical systems theory. We begin with principles for designing finite and infinitesimal motions in small modules, and show that each module is a one-dimensional map between distances across pairs of nodes. Next, we represent the act of combining modules as an iteration of this map, and design networks whose geometries reflect the map's fixed points, limit cycles, and chaos. We use this representation to design different folding sequences from a deployable network and a soliton, to a branched network acting as a mechanical AND gate. Finally, we design large changes in curvature of the entire network, and construct physical networks from laser-cut acrylic, origami, and 3D printed material to demonstrate the framework's potential and versatility for designing the full conformational trajectory of morphing metamaterials and structures.

cond-mat.soft↗

arXiv2019

Linear Dynamics & Control of Brain Networks

The brain is an intricately structured organ responsible for the rich emergent dynamics that support the complex cognitive functions we enjoy as humans. With around $10^{11}$ neurons and $10^{15}$ synapses, understanding how the human brain works has proven to be a daunting endeavor, requiring concerted collaboration across traditional disciplinary boundaries. In some cases, that collaboration has occurred between experimentalists and technicians, who offer new physical tools to measure and manipulate neural function. In other contexts, that collaboration has occurred between experimentalists and theorists, who offer new conceptual tools to explain existing data and inform new directions for empirical research. In this chapter, we offer an example of the latter. Specifically, we focus on the simple but powerful framework of linear systems theory as a useful tool both for capturing biophysically relevant parameters of neural activity and connectivity, and for analytical and numerical study. We begin with a brief overview of state-space representations and linearization of neural models for non-linear dynamical systems. We then derive core concepts in the theory of linear systems such as the impulse and controlled responses to external stimuli, achieving desired state transitions, controllability, and minimum energy control. Afterwards, we discuss recent advances in the application of linear systems theory to structural and functional brain data across multiple spatial and temporal scales, along with methodological considerations and limitations. We close with a brief discussion of open frontiers and our vision for the future.

q-bio.NC↗

arXiv2018

Conformational Control of Mechanical Networks

Understanding conformational change is crucial for programming and controlling the function of many mechanical systems such as allosteric enzymes and tunable metamaterials. Of particular interest is the relationship between the network topology or geometry and the specific motions observed under controlling perturbations. We study this relationship in mechanical networks of 2-D and 3-D Maxwell frames composed of point masses connected by rigid rods rotating freely about the masses. We first develop simple principles that yield all bipartite network topologies and geometries that give rise to an arbitrarily specified instantaneous and finitely deformable motion in the masses as the sole non-rigid body zero mode. We then extend these principles to characterize networks that simultaneously yield multiple specified zero modes, and create large networks by coupling individual modules. These principles are then used to characterize and design networks with useful material (negative Poisson ratio) and mechanical (targeted allosteric response) functions.

physics.comp-ph↗