arXiv subjects

Ankush Chakrabarty

Publications and source records attributed to Ankush Chakrabarty.

At least 19 recordsLinked to original sources

arXiv2026

Training Language Models to Cooperate with Inference-Time Controllers

Large language model (LLM) performance increasingly depends not only on the base model, but also on the inference-time controller used to organize reasoning. Existing post-training methods, however, typically optimize for a single fixed interaction pattern, despite real deployments relying on diverse controllers such as Chain-of-Thought, self-consistency, debate, planning, and verification pipelines. This creates a training--deployment mismatch and limits transfer to new workflows. We introduce CALM (Controller-Aware Language Models), a post-training framework that explicitly places controllers in the training loop. We formulate controller-aware post-training as multi-task reinforcement learning over controller-induced interaction protocols, where controllers are compositions of reusable local reasoning modules. This structure also induces a module-level decomposition of mixed-controller training under a turn-level GRPO objective, enabling a systematic study of controller and module-aware training strategies. We evaluate CALM on held-out controller compositions and broader controller shifts, showing that controller-aware post-training improves generalization across inference-time workflows beyond single-controller optimization.

cs.AI↗

arXiv2026

Meta-Learning for Rapid Adaptation in Reference Tracking of Uncertain Nonlinear Systems

In this paper, we address the problem of reference tracking for uncertain nonlinear systems. Since collecting data from the target system (i.e., the system of interest) is often challenging, our objective is to design optimal controllers using limited target system data. Meta-learning provides a promising paradigm by leveraging offline data from source systems (systems sharing structural similarities with the target system) to accelerate training and enhance control performance. Motivated by this idea, we propose a meta-learning-based control framework that tailors the implicit model-agnostic meta-learning (iMAML) algorithm to the control setting. The framework operates in two phases: an (offline) meta-training phase, where an aggregated representation is learned from source data to capture the shared system dynamics among similar systems, and an (online) meta-adaptation phase, where this representation is fine-tuned on the target system using only a few data samples and limited adaptation steps. We formulate this framework as a bi-level optimization problem and provide an efficient solution with reduced storage complexity and few approximations. The proposed framework is general, allowing various learning algorithms to be integrated. To demonstrate this flexibility, we propose two specific learning algorithms that can be incorporated into our framework based on a neural state-space model and a deep Q-network, respectively. The primary distinction between these approaches is whether explicit system identification is required. Numerical simulations and hardware experiments demonstrate that the proposed methods enhance control performance and consistently outperform baseline approaches.

cs.AI↗

arXiv2025

Probabilistic Forecasting for Building Energy Systems using Time-Series Foundation Models

Decision-making in building energy systems critically depends on the predictive accuracy of relevant time-series models. In scenarios lacking extensive data from a target building, foundation models (FMs) represent a promising technology that can leverage prior knowledge from vast and diverse pre-training datasets to construct accurate probabilistic predictors for use in decision-making tools. This paper investigates the applicability and fine-tuning strategies of time-series foundation models (TSFMs) in building energy forecasting. We analyze both full fine-tuning and parameter-efficient fine-tuning approaches, particularly low-rank adaptation (LoRA), by using real-world data from a commercial net-zero energy building to capture signals such as room occupancy, carbon emissions, plug loads, and HVAC energy consumption. Our analysis reveals that the zero-shot predictive performance of TSFMs is generally suboptimal. To address this shortcoming, we demonstrate that employing either full fine-tuning or parameter-efficient fine-tuning significantly enhances forecasting accuracy, even with limited historical data. Notably, fine-tuning with low-rank adaptation (LoRA) substantially reduces computational costs without sacrificing accuracy. Furthermore, fine-tuned TSFMs consistently outperform state-of-the-art deep forecasting models (e.g., temporal fusion transformers) in accuracy, robustness, and generalization across varying building zones and seasonal conditions. These results underline the efficacy of TSFMs for practical, data-constrained building energy management systems, enabling improved decision-making in pursuit of energy efficiency and sustainability.

cs.LG↗

arXiv2025

Manifold meta-learning for reduced-complexity neural system identification

System identification has greatly benefited from deep learning techniques, particularly for modeling complex, nonlinear dynamical systems with partially unknown physics where traditional approaches may not be feasible. However, deep learning models often require large datasets and significant computational resources at training and inference due to their high-dimensional parameterizations. To address this challenge, we propose a meta-learning framework that discovers a low-dimensional manifold within the parameter space of an over-parameterized neural network architecture. This manifold is learned from a meta-dataset of input-output sequences generated by a class of related dynamical systems, enabling efficient model training while preserving the network's expressive power for the considered system class. Unlike bilevel meta-learning approaches, our method employs an auxiliary neural network to map datasets directly onto the learned manifold, eliminating the need for costly second-order gradient computations during meta-training and reducing the number of first-order updates required in inference, which could be expensive for large models. We validate our approach on a family of Bouc-Wen oscillators, which is a well-studied nonlinear system identification benchmark. We demonstrate that we are able to learn accurate models even in small-data scenarios.

cs.LG↗

arXiv2025

User Preference Meets Pareto-Optimality in Multi-Objective Bayesian Optimization

Incorporating user preferences into multi-objective Bayesian optimization (MOBO) allows for personalization of the optimization procedure. Preferences are often abstracted in the form of an unknown utility function, estimated through pairwise comparisons of potential outcomes. However, utility-driven MOBO methods can yield solutions that are dominated by nearby solutions, as non-dominance is not enforced. Additionally, classical MOBO commonly relies on estimating the entire Pareto-front to identify the Pareto-optimal solutions, which can be expensive and ignore user preferences. Here, we present a new method, termed preference-utility-balanced MOBO (PUB-MOBO), that allows users to disambiguate between near-Pareto candidate solutions. PUB-MOBO combines utility-based MOBO with local multi-gradient descent to refine user-preferred solutions to be near-Pareto-optimal. To this end, we propose a novel preference-dominated utility function that concurrently preserves user-preferences and dominance amongst candidate solutions. A key advantage of PUB-MOBO is that the local search is restricted to a (small) region of the Pareto-front directed by user preferences, alleviating the need to estimate the entire Pareto-front. PUB-MOBO is tested on three synthetic benchmark problems: DTLZ1, DTLZ2 and DH1, as well as on three real-world problems: Vehicle Safety, Conceptual Marine Design, and Car Side Impact. PUB-MOBO consistently outperforms state-of-the-art competitors in terms of proximity to the Pareto-front and utility regret across all the problems.

cs.LG↗

arXiv2025

Meta-Learning for Physically-Constrained Neural System Identification

We present a gradient-based meta-learning framework for rapid adaptation of neural state-space models (NSSMs) for black-box system identification. When applicable, we also incorporate domain-specific physical constraints to improve the accuracy of the NSSM. The major benefit of our approach is that instead of relying solely on data from a single target system, our framework utilizes data from a diverse set of source systems, enabling learning from limited target data, as well as with few online training iterations. Through benchmark examples, we demonstrate the potential of our approach, study the effect of fine-tuning subnetworks rather than full fine-tuning, and report real-world case studies to illustrate the practical application and generalizability of the approach to practical problems with physical-constraints. Specifically, we show that the meta-learned models result in improved downstream performance in model-based state estimation in indoor localization and energy systems.

cs.LG↗

arXiv2024

BEACON: A Bayesian Optimization Inspired Strategy for Efficient Novelty Search

Novelty search (NS) aims to uncover diverse system behaviors through simulation or experiment without requiring a pre-specified scalar objective. This capability is especially relevant to modern discovery problems in chemistry, materials science, and molecular design, where researchers often seek broad coverage of attainable property space rather than a single optimum and where each evaluation may require a costly computation or experiment. For such expensive black-box settings, we propose BEACON, a sample-efficient NS strategy inspired by Bayesian optimization principles. BEACON models the input-to-outcome mapping using multi-output Gaussian processes and selects new inputs by scoring how far plausible posterior outcomes lie from a denoised archive of previously observed outcomes. This gives a distance-based novelty acquisition that accounts for predictive uncertainty and observational noise while operating directly in continuous outcome space, rather than requiring direct optimization over a discretized partition of behaviors. By leveraging efficient posterior sampling together with scalable high-dimensional Gaussian process models, the proposed framework can be extended to settings with large data sets and high-dimensional design variables. We demonstrate BEACON on established benchmark problems together with real-world case studies in materials and molecular discovery. Across these settings, BEACON consistently discovers broader sets of distinct behaviors than several competing baselines under limited evaluation budgets.

stat.ML↗

arXiv2024

MPC of Uncertain Nonlinear Systems with Meta-Learning for Fast Adaptation of Neural Predictive Models

In this paper, we consider the problem of reference tracking in uncertain nonlinear systems. A neural State-Space Model (NSSM) is used to approximate the nonlinear system, where a deep encoder network learns the nonlinearity from data, and a state-space component captures the temporal relationship. This transforms the nonlinear system into a linear system in a latent space, enabling the application of model predictive control (MPC) to determine effective control actions. Our objective is to design the optimal controller using limited data from the \textit{target system} (the system of interest). To this end, we employ an implicit model-agnostic meta-learning (iMAML) framework that leverages information from \textit{source systems} (systems that share similarities with the target system) to expedite training in the target system and enhance its control performance. The framework consists of two phases: the (offine) meta-training phase learns a aggregated NSSM using data from source systems, and the (online) meta-inference phase quickly adapts this aggregated model to the target system using only a few data points and few online training iterations, based on local loss function gradients. The iMAML algorithm exploits the implicit function theorem to exactly compute the gradient during training, without relying on the entire optimization path. By focusing solely on the optimal solution, rather than the path, we can meta-train with less storage complexity and fewer approximations than other contemporary meta-learning algorithms. We demonstrate through numerical examples that our proposed method can yield accurate predictive models by adaptation, resulting in a downstream MPC that outperforms several baselines.

eess.SY↗

arXiv2023

Safe multi-agent motion planning under uncertainty for drones using filtered reinforcement learning

We consider the problem of safe multi-agent motion planning for drones in uncertain, cluttered workspaces. For this problem, we present a tractable motion planner that builds upon the strengths of reinforcement learning and constrained-control-based trajectory planning. First, we use single-agent reinforcement learning to learn motion plans from data that reach the target but may not be collision-free. Next, we use a convex optimization, chance constraints, and set-based methods for constrained control to ensure safety, despite the uncertainty in the workspace, agent motion, and sensing. The proposed approach can handle state and control constraints on the agents, and enforce collision avoidance among themselves and with static obstacles in the workspace with high probability. The proposed approach yields a safe, real-time implementable, multi-agent motion planner that is simpler to train than methods based solely on learning. Numerical simulations and experiments show the efficacy of the approach.

cs.RO↗

arXiv2023

Physics-Informed Machine Learning for Modeling and Control of Dynamical Systems

Physics-informed machine learning (PIML) is a set of methods and tools that systematically integrate machine learning (ML) algorithms with physical constraints and abstract mathematical models developed in scientific and engineering domains. As opposed to purely data-driven methods, PIML models can be trained from additional information obtained by enforcing physical laws such as energy and mass conservation. More broadly, PIML models can include abstract properties and conditions such as stability, convexity, or invariance. The basic premise of PIML is that the integration of ML and physics can yield more effective, physically consistent, and data-efficient models. This paper aims to provide a tutorial-like overview of the recent advances in PIML for dynamical system modeling and control. Specifically, the paper covers an overview of the theory, fundamental concepts and methods, tools, and applications on topics of: 1) physics-informed learning for system identification; 2) physics-informed learning for control; 3) analysis and verification of PIML models; and 4) physics-informed digital twins. The paper is concluded with a perspective on open challenges and future research opportunities.

eess.SY↗

arXiv2023

Violation-Aware Contextual Bayesian Optimization for Controller Performance Optimization with Unmodeled Constraints

We study the problem of performance optimization of closed-loop control systems with unmodeled dynamics. Bayesian optimization (BO) has been demonstrated to be effective for improving closed-loop performance by automatically tuning controller gains or reference setpoints in a model-free manner. However, BO methods have rarely been tested on dynamical systems with unmodeled constraints and time-varying ambient conditions. In this paper, we propose a violation-aware contextual BO algorithm (VACBO) that optimizes closed-loop performance while simultaneously learning constraint-feasible solutions under time-varying ambient conditions. Unlike classical constrained BO methods which allow unlimited constraint violations, or 'safe' BO algorithms that are conservative and try to operate with near-zero violations, we allow budgeted constraint violations to improve constraint learning and accelerate optimization. We demonstrate the effectiveness of our proposed VACBO method for energy minimization of industrial vapor compression systems under time-varying ambient temperature and humidity.

cs.LG↗

arXiv2023

Data-Driven Identification of Dynamic Quality Models in Drinking Water Networks

Traditional control and monitoring of water quality in drinking water distribution networks (WDN) rely on mostly model- or toolbox-driven approaches, where the network topology and parameters are assumed to be known. In contrast, system identification (SysID) algorithms for generic dynamic system models seek to approximate such models using only input-output data without relying on network parameters. The objective of this paper is to investigate SysID algorithms for water quality model approximation. This research problem is challenging due to (i) complex water quality and reaction dynamics and (ii) the mismatch between the requirements of SysID algorithms and the properties of water quality dynamics. In this paper, we present the first attempt to identify water quality models in WDNs using only input-output experimental data and classical SysID methods without knowing any WDN parameters. Properties of water quality models are introduced, the ensuing challenges caused by these properties when identifying water quality models are discussed, and remedial solutions are given. Through case studies, we demonstrate the applicability of SysID algorithms, show the corresponding performance in terms of accuracy and computational time, and explore the possible factors impacting water quality model identification.

eess.SY↗

arXiv2022

Meta-Learning of Neural State-Space Models Using Data From Similar Systems

Deep neural state-space models (SSMs) provide a powerful tool for modeling dynamical systems solely using operational data. Typically, neural SSMs are trained using data collected from the actual system under consideration, despite the likely existence of operational data from similar systems which have previously been deployed in the field. In this paper, we propose the use of model-agnostic meta-learning (MAML) for constructing deep encoder network-based SSMs, by leveraging a combination of archived data from similar systems (used to meta-train offline) and limited data from the actual system (used for rapid online adaptation). We demonstrate using a numerical example that meta-learning can result in more accurate neural SSM models than supervised- or transfer-learning, despite few adaptation steps and limited online data. Additionally, we show that by carefully partitioning and adapting the encoder layers while fixing the state-transition operator, we can achieve comparable performance to MAML while reducing online adaptation complexity.

cs.LG↗

arXiv2022

Optimizing Closed-Loop Performance with Data from Similar Systems: A Bayesian Meta-Learning Approach

Bayesian optimization (BO) has demonstrated potential for optimizing control performance in data-limited settings, especially for systems with unknown dynamics or unmodeled performance objectives. The BO algorithm efficiently trades-off exploration and exploitation by leveraging uncertainty estimates using surrogate models. These surrogates are usually learned using data collected from the target dynamical system to be optimized. Intuitively, the convergence rate of BO is better for surrogate models that can accurately predict the target system performance. In classical BO, initial surrogate models are constructed using very limited data points, and therefore rarely yield accurate predictions of system performance. In this paper, we propose the use of meta-learning to generate an initial surrogate model based on data collected from performance optimization tasks performed on a variety of systems that are different to the target system. To this end, we employ deep kernel networks (DKNs) which are simple to train and which comprise encoded Gaussian process models that integrate seamlessly with classical BO. The effectiveness of our proposed DKN-BO approach for speeding up control system performance optimization is demonstrated using a well-studied nonlinear system with unknown dynamics and an unmodeled performance function.

cs.LG↗

arXiv2022

Model Order Reduction for Water Quality Dynamics

A state-space representation of water quality dynamics describing disinfectant (e.g., chlorine) transport dynamics in drinking water distribution networks has been recently proposed. Such representation is a byproduct of space- and time-discretization of the PDE modeling transport dynamics. This results in a large state-space dimension even for small networks with tens of nodes. Although such a state-space model provides a model-driven approach to predict water quality dynamics, incorporating it into model-based control algorithms or state estimators for large networks is challenging and at times intractable. To that end, this paper investigates model order reduction (MOR) methods for water quality dynamics with the objective of performing post-reduction feedback control. The presented investigation focuses on reducing state-dimension by orders of magnitude, the stability of the MOR methods, and the application of these methods to model predictive control.

eess.SY↗

arXiv2021

Extremum Seeking Control with an Adaptive Gain Based On Gradient Estimation Error

This paper presents an extremum seeking control algorithm with an adaptive step-size that adjusts the aggressiveness of the controller based on the quality of the gradient estimate. The adaptive step-size ensures that the integral-action produced by the gradient descent does not destabilize the closed-loop system. To quantify the quality of the gradient estimate, we present a batch least squares estimator with a novel weighting and show that it produces bounded estimation errors, where the uncertainty is due to the curvature of the unknown cost function. The adaptive step-size then maximizes the decrease of the combined plant and controller Lyapunov function for the worst-case estimation error. We prove that our ESC is input-to-state stable with respect to the dither signal. Finally, we demonstrate our proposed ESC through five numerical examples; one illustrative, one practical, and three benchmarks.

math.OC↗

arXiv2021

VABO: Violation-Aware Bayesian Optimization for Closed-Loop Control Performance Optimization with Unmodeled Constraints

We study the problem of performance optimization of closed-loop control systems with unmodeled dynamics. Bayesian optimization (BO) has been demonstrated effective for improving closed-loop performance by automatically tuning controller gains or reference setpoints in a model-free manner. However, BO methods have rarely been tested on dynamical systems with unmodeled constraints. In this paper, we propose a violation-aware BO algorithm (VABO) that optimizes closed-loop performance while simultaneously learning constraint-feasible solutions. Unlike classical constrained BO methods which allow an unlimited constraint violations, or safe BO algorithms that are conservative and try to operate with near-zero violations, we allow budgeted constraint violations to improve constraint learning and accelerate optimization. We demonstrate the effectiveness of our proposed VABO method for energy minimization of industrial vapor compression systems.

cs.LG↗

arXiv2021

Attentive Neural Processes and Batch Bayesian Optimization for Scalable Calibration of Physics-Informed Digital Twins

Physics-informed dynamical system models form critical components of digital twins of the built environment. These digital twins enable the design of energy-efficient infrastructure, but must be properly calibrated to accurately reflect system behavior for downstream prediction and analysis. Dynamical system models of modern buildings are typically described by a large number of parameters and incur significant computational expenditure during simulations. To handle large-scale calibration of digital twins without exorbitant simulations, we propose ANP-BBO: a scalable and parallelizable batch-wise Bayesian optimization (BBO) methodology that leverages attentive neural processes (ANPs).

cs.LG↗