arXiv subjects

Fabio Pasqualetti

Publications and source records attributed to Fabio Pasqualetti.

At least 55 records · Page 3Linked to original sources

arXiv2023

Learning on Manifolds: Universal Approximations Properties using Geometric Controllability Conditions for Neural ODEs

In numerous robotics and mechanical engineering applications, among others, data is often constrained on smooth manifolds due to the presence of rotational degrees of freedom. Common datadriven and learning-based methods such as neural ordinary differential equations (ODEs), however, typically fail to satisfy these manifold constraints and perform poorly for these applications. To address this shortcoming, in this paper we study a class of neural ordinary differential equations that, by design, leave a given manifold invariant, and characterize their properties by leveraging the controllability properties of control affine systems. In particular, using a result due to Agrachev and Caponigro on approximating diffeomorphisms with flows of feedback control systems, we show that any map that can be represented as the flow of a manifold-constrained dynamical system can also be approximated using the flow of manifold-constrained neural ODE, whenever a certain controllability condition is satisfied. Additionally, we show that this universal approximation property holds when the neural ODE has limited width in each layer, thus leveraging the depth of network instead for approximation. We verify our theoretical findings using numerical experiments on PyTorch for the manifolds S2 and the 3-dimensional orthogonal group SO(3), which are model manifolds for mechanical systems such as spacecrafts and satellites. We also compare the performance of the manifold invariant neural ODE with classical neural ODEs that ignore the manifold invariant properties and show the superiority of our approach in terms of accuracy and sample complexity.

math.OC↗

arXiv2023

Model-based and Data-based Dynamic Output Feedback for Externally Positive Systems

In this work, we derive dynamic output-feedback controllers that render the closed-loop system externally positive. We begin by expressing the class of discrete-time, linear, time-invariant systems and the class of dynamic controllers in the space of input-output behaviors, where a dynamic controller can be expressed as a static behavioral feedback gain. We leverage the static form of the controller to derive output-feedback controllers that achieve monotonic output tracking of a constant non-negative reference output. Further, we provide a direct data-driven approach to derive monotonic tracking output-feedback controllers for single-input-single-output (SISO) systems. Our approaches, model-based and data-based, allow us to obtain output-feedback controllers that render the closed-loop system externally positive. Finally, we validate our results numerically in a drone landing control problem.

eess.SY↗

arXiv2023

Data-driven Eigenstructure Assignment for Sparse Feedback Design

This paper presents a novel approach for solving the pole placement and eigenstructure assignment problems through data-driven methods. By using open-loop data alone, the paper shows that it is possible to characterize the allowable eigenvector subspaces, as well as the set of feedback gains that solve the pole placement problem. Additionally, the paper proposes a closed-form expression for the feedback gain that solves the eigenstructure assignment problem. Finally, the paper discusses a series of optimization problems aimed at finding sparse feedback gains for the pole placement problem.

eess.SY↗

arXiv2023

Stochastic Contextual Bandits with Long Horizon Rewards

The growing interest in complex decision-making and language modeling problems highlights the importance of sample-efficient learning over very long horizons. This work takes a step in this direction by investigating contextual linear bandits where the current reward depends on at most $s$ prior actions and contexts (not necessarily consecutive), up to a time horizon of $h$. In order to avoid polynomial dependence on $h$, we propose new algorithms that leverage sparsity to discover the dependence pattern and arm parameters jointly. We consider both the data-poor ($T<h$) and data-rich ($T\ge h$) regimes, and derive respective regret upper bounds $\tilde O(d\sqrt{sT} +\min\{ q, T\})$ and $\tilde O(\sqrt{sdT})$, with sparsity $s$, feature dimension $d$, total time horizon $T$, and $q$ that is adaptive to the reward dependence pattern. Complementing upper bounds, we also show that learning over a single trajectory brings inherent challenges: While the dependence pattern and arm parameters form a rank-1 matrix, circulant matrices are not isometric over rank-1 manifolds and sample complexity indeed benefits from the sparse reward dependence structure. Our results necessitate a new analysis to address long-range temporal dependencies across data and avoid polynomial dependence on the reward horizon $h$. Specifically, we utilize connections to the restricted isometry property of circulant matrices formed by dependent sub-Gaussian vectors and establish new guarantees that are also of independent interest.

cs.LG↗

arXiv2022

Behavioral Feedback for Optimal LQG Control

In this work, we revisit the Linear Quadratic Gaussian (LQG) optimal control problem from a behavioral perspective. Motivated by the suitability of behavioral models for data-driven control, we begin with a reformulation of the LQG problem in the space of input-output behaviors and obtain a complete characterization of the optimal solutions. In particular, we show that the optimal LQG controller can be expressed as a static behavioral-feedback gain, thereby eliminating the need for dynamic state estimation characteristic of state space methods. The static form of the optimal LQG gain also makes it amenable to its computation by gradient descent, which we investigate via numerical experiments. Furthermore, we highlight the advantage of this approach in the data-driven control setting of learning the optimal LQG controller from expert demonstrations.

eess.SY↗

arXiv2022

Direct vs Indirect Methods for Behavior-based Attack Detection

We study the problem of data-driven attack detection for unknown LTI systems using only input-output behavioral data. In contrast with model-based detectors that use errors from an output predictor to detect attacks, we study behavior-based data-driven detectors. We construct a behavior-based chi-squared detector that uses a sequence of inputs and outputs and their covariance. The covariance of the behaviors is estimated using data by two methods. The first (direct) method employs the sample covariance as an estimate of the covariance of behaviors. The second (indirect) method uses a lower dimensional generative model identified from data to estimate the covariance of behaviors. We prove the consistency of the two methods of estimation and provide finite sample error bounds. Finally, we numerically compare the performance and establish a tradeoff between the methods at different regimes of the size of the data set and the length of the detection horizon. Our numerical study indicates that neither method is invariable superior, and reveals the existence of two regimes for the performance of the two methods, wherein the direct method is superior in cases with large data sets relative to the length of the detection horizon, while the indirect method is superior in cases with small data sets.

math.OC↗

arXiv2022

Representation Learning for Context-Dependent Decision-Making

Humans are capable of adjusting to changing environments flexibly and quickly. Empirical evidence has revealed that representation learning plays a crucial role in endowing humans with such a capability. Inspired by this observation, we study representation learning in the sequential decision-making scenario with contextual changes. We propose an online algorithm that is able to learn and transfer context-dependent representations and show that it significantly outperforms the existing ones that do not learn representations adaptively. As a case study, we apply our algorithm to the Wisconsin Card Sorting Task, a well-established test for the mental flexibility of humans in sequential decision-making. By comparing our algorithm with the standard Q-learning and Deep-Q learning algorithms, we demonstrate the benefits of adaptive representation learning.

cs.LG↗

arXiv2022

Non-Stationary Representation Learning in Sequential Linear Bandits

In this paper, we study representation learning for multi-task decision-making in non-stationary environments. We consider the framework of sequential linear bandits, where the agent performs a series of tasks drawn from distinct sets associated with different environments. The embeddings of tasks in each set share a low-dimensional feature extractor called representation, and representations are different across sets. We propose an online algorithm that facilitates efficient decision-making by learning and transferring non-stationary representations in an adaptive fashion. We prove that our algorithm significantly outperforms the existing ones that treat tasks independently. We also conduct experiments using both synthetic and real data to validate our theoretical insights and demonstrate the efficacy of our algorithm.

cs.LG↗

arXiv2022

Vibrational Control of Cluster Synchronization: Connections with Deep Brain Stimulation

Cluster synchronization underlies various functions in the brain. Abnormal patterns of cluster synchronization are often associated with neurological disorders. Deep brain stimulation (DBS) is a neurosurgical technique used to treat several brain diseases, which has been observed to regulate neuronal synchrony patterns. Despite its widespread use, the mechanisms of DBS remain largely unknown. In this paper, we hypothesize that DBS plays a role similar to vibrational control since they both highly rely on high-frequency excitation to function. Under the framework of Kuramoto-oscillator networks, we study how vibrations introduced to network connections can stabilize cluster synchronization. We derive some sufficient conditions and also provide an effective approach to design vibrational control. Also, a numerical example is presented to demonstrate our theoretical findings.

eess.SY↗

arXiv2022

Parameter Conditions to Prevent Voltage Oscillations Caused by LTC-Inverter Hunting on Power Distribution Grids

As more distributed energy resources (DERs) are connected to the power grid, it becomes increasingly important to ensure safe and effective coordination between legacy voltage regulation devices and inverter-based DERs. In this work, we show how a distribution circuit model, composed of two LTCs and two inverter devices, can create voltage oscillations even with reasonable choices of control parameters. By modeling the four-device circuit as a switched affine hybrid system, we analyze the system's oscillatory behavior, both during normal operation and after a cyber-physical attack. Through the analysis we determine the specific region of the voltage state space where oscillations are possible and derive conditions on the control parameters to guarantee against the oscillations. Finally, we project the derived parameter conditions onto 2D spaces, and describe the application of our problem formulation to grids with many devices.

eess.SY↗

arXiv2022

Functional Control of Oscillator Networks

Oscillatory activity is ubiquitous in natural and engineered network systems. The interaction scheme underlying interdependent oscillatory components governs the emergence of network-wide patterns of synchrony that regulate and enable complex functions. Yet, understanding, and ultimately harnessing, the structure-function relationship in oscillator networks remains an outstanding challenge of modern science. Here, we address this challenge by presenting a principled method to prescribe exact and robust functional configurations from local network interactions through optimal tuning of the oscillators' parameters. To quantify the behavioral synchrony between coupled oscillators, we introduce the notion of functional pattern, which encodes the pairwise relationships between the oscillators' phases. Our procedure is computationally efficient and provably correct, accounts for constrained interaction types, and allows to concurrently assign multiple desired functional patterns. Further, we derive algebraic and graph-theoretic conditions to guarantee the feasibility and stability of target functional patterns. These conditions provide an interpretable mapping between the structural constraints and their functional implications in oscillator networks. As a proof of concept, we apply the proposed method to replicate empirically recorded functional relationships from cortical oscillations in a human brain, and to redistribute the active power flow in different models of electrical grids.

nlin.AO↗

arXiv2022

Learning Lipschitz Feedback Policies from Expert Demonstrations: Closed-Loop Guarantees, Generalization and Robustness

In this work, we propose a framework to learn feedback control policies with guarantees on closed-loop generalization and adversarial robustness. These policies are learned directly from expert demonstrations, contained in a dataset of state-control input pairs, without any prior knowledge of the task and system model. We use a Lipschitz-constrained loss minimization scheme to learn feedback policies with certified closed-loop robustness, wherein the Lipschitz constraint serves as a mechanism to tune the generalization performance and robustness to adversarial disturbances. Our analysis exploits the Lipschitz property to obtain closed-loop guarantees on generalization and robustness of the learned policies. In particular, we derive a finite sample bound on the policy learning error and establish robust closed-loop stability under the learned control policy. We also derive bounds on the closed-loop regret with respect to the expert policy and the deterioration of closed-loop performance under bounded (adversarial) disturbances to the state measurements. Numerical results validate our analysis and demonstrate the effectiveness of our robust feedback policy learning framework. Finally, our results suggest the existence of a potential tradeoff between nominal closed-loop performance and adversarial robustness, and that improvements in nominal closed-loop performance can only be made at the expense of robustness to adversarial perturbations.

cs.LG↗

arXiv2022

Data-driven Meets Geometric Control: Zero Dynamics, Subspace Stabilization, and Malicious Attacks

Studying structural properties of linear dynamical systems through invariant subspaces is one of the key contributions of the geometric approach to system theory. In general, a model of the dynamics is required in order to compute the invariant subspaces of interest. In this paper we overcome this limitation by finding data-driven formulas for some of the foundational tools of geometric control. In particular, for an unknown linear system, we show how controlled and conditioned invariant subspaces can be found directly from experimental data. We use our formulas and approach to (i) find a feedback gain that confines the system state within a desired subspace, (ii) compute the invariant zeros of the unknown system, and (iii) design attacks that remain undetectable.

eess.SY↗

arXiv2021

Robust Adversarial Classification via Abstaining

In this work, we consider a binary classification problem and cast it into a binary hypothesis testing framework, where the observations can be perturbed by an adversary. To improve the adversarial robustness of a classifier, we include an abstain option, where the classifier abstains from making a decision when it has low confidence about the prediction. We propose metrics to quantify the nominal performance of a classifier with an abstain option and its robustness against adversarial perturbations. We show that there exist a tradeoff between the two metrics regardless of what method is used to choose the abstain region. Our results imply that the robustness of a classifier with an abstain option can only be improved at the expense of its nominal performance. Further, we provide necessary conditions to design the abstain region for a 1- dimensional binary classification problem. We validate our theoretical results on the MNIST dataset, where we numerically show that the tradeoff between performance and robustness also exist for the general multi-class classification problems.

cs.LG↗

arXiv2021

On Direct vs Indirect Data-Driven Predictive Control

In this work, we compare the direct and indirect approaches to data-driven predictive control of stochastic linear time-invariant systems. The distinction between the two approaches lies in the fact that the indirect approach involves identifying a lower dimensional model from data which is then used in a certainty-equivalent control design, while the direct approach avoids this intermediate step altogether. Working within an optimization-based framework, we find that the suboptimality gap measuring the control performance w.r.t. the optimal model-based control design vanishes with the size of the dataset only with the direct approach. The indirect approach has a higher rate of convergence, but its suboptimality gap does not vanish as the size of the dataset increases. This reveals the existence of two distinct regimes of performance as the size of the dataset of input-output behaviors is increased. We show that the indirect approach, by relying on the identification of a lower dimensional model, has lower variance and outperforms the direct approach for smaller datasets, while it incurs an asymptotic bias as a result of the process noise and a (possibly) incorrect assumption on the order of the identified model. The direct approach, however, does not incur an asymptotic bias, and outperforms the indirect approach for larger datasets. Ultimately, by revealing the existence of two non-asymptotic regimes for the performance of direct and indirect data-driven predictive control designs, our study suggests that neither approach is invariably superior and that the choice of design must, in practice, be informed by the available dataset.

math.OC↗

arXiv2021

Structural underpinnings of control in multiplex networks

To design control strategies that predictably manipulate a system's behavior, it is first necessary to understand how the system's structure relates to its response. Many complex systems can be represented as multilayer networks whose response and control can be studied in the framework of network control theory. It remains unknown how a system's layered architecture dictates its control properties, particularly when control signals can only access the system through a single input layer. Here, we use the framework of linear control theory to probe the control properties of a duplex network with directed interlayer links. We determine the manner in which the structural properties of layers and relative interlayer arrangement together dictate the system's response. For this purpose, we calculate the exact expression of optimal control energy in terms of layer spectra and the relative alignment between the eigenmodes of the input layer and the deeper target layer. For a range of numerically constructed duplex networks, we then calculate the control properties of the two layers as a function of target-layer densities and layer topology. The alignment of layer eigenmodes emerges as an important parameter that sets the cost and routing of the optimal energy for control. We understand these results in a simplified limit of a single-mode approximation, and we build metrics to characterize the routing of optimal control energy through the eigenmodes of each layer. Our analytical and numerical results together provide insights into the relationships between structure and control in duplex networks. More generally, they serve as a platform for future work designing optimal \emph{interlayer} control strategies.

q-bio.NC↗

arXiv2020

Path-dependent connectivity, not modularity, consistently predicts controllability of structural brain networks

The human brain displays rich communication dynamics that are thought to be particularly well-reflected in its marked community structure. Yet, the precise relationship between community structure in structural brain networks and the communication dynamics that can emerge therefrom is not well-understood. In addition to offering insight into the structure-function relationship of networked systems, such an understanding is a critical step towards the ability to manipulate the brain's large-scale dynamical activity in a targeted manner. We investigate the role of community structure in the controllability of structural brain networks. At the region level, we find that certain network measures of community structure are sometimes statistically correlated with measures of linear controllability. However, we then demonstrate that this relationship depends on the distribution of network edge weights. We highlight the complexity of the relationship between community structure and controllability by performing numerical simulations using canonical graph models with varying mesoscale architectures and edge weight distributions. Finally, we demonstrate that weighted subgraph centrality, a measure rooted in the graph spectrum, and which captures higher-order graph architecture, is a stronger and more consistent predictor of controllability. Our study contributes to an understanding of how the brain's diverse mesoscale structure supports transient communication dynamics.

q-bio.NC↗

arXiv2020

Phase-Amplitude Coupling in Neuronal Oscillator Networks

Phase-amplitude coupling (PAC) describes the phenomenon where the power of a high-frequency oscillation evolves with the phase of a low-frequency one. We propose a model that explains the emergence of PAC in two commonly-accepted architectures in the brain, namely, a high-frequency neural oscillation driven by an external low-frequency input and two interacting local oscillations with distinct, locally-generated frequencies. We further propose an interconnection structure for brain regions and demonstrate that low-frequency phase synchrony can integrate high-frequency activities regulated by local PAC and control the direction of information flow across distant regions.

q-bio.NC↗