Searcharxiv⌕ Search

arXiv subjects

Parikshit Pareek

Publications and source records attributed to Parikshit Pareek.

At least 19 recordsLinked to original sources

Generalization Testing of Grid Foundation Models for Power Flow and Optimal Power Flow

There is growing interest in foundation models (FMs) for power grids, particularly for power flow and optimal power flow. Their usefulness depends on generalization accuracy, which existing works test on their own benchmarks. We propose a three-axis framework to test their generalization limits: changed admittance matrices of training grids, load levels, and unseen grids. We test GENCO, LUMINA and GridSFM against the DC baseline, alongside their reported limitations, and discuss theoretical reasons for their failures. Experiments reveal that adding a line degrades all FMs more than removing one, and longer lines eventually make LUMINA and fine-tuned GENCO worse than DC. All three degrade beyond their training load range, where the loading-voltage curve steepens. On unseen grids, LUMINA degrades sharply even inside its training range, and GridSFM fails on the grid farthest from it. Thus, beyond training conditions, FMs behave more like meta-models requiring test-time fine-tuning data and compute.

eess.SY↗

State Estimation as In-context Tabular Prediction

This letter poses state estimation (SE) as an in-context tabular prediction (ICTP) problem. The key idea is a tabular arrangement in which historical measurement vectors and corresponding state estimates serve as the context, and a new measurement vector forms the query for which we predict a state vector. We propose to use tabular foundation models (TFMs), pretrained on synthetic data, whose weights are never updated, so power system data is needed only as context. With frozen TFM weights we estimate the state of 14-, 30-, 33-, 70-, and 118-bus transmission and distribution networks, and do so where weighted least squares (WLS) fails or degrades, under low observability and with bad data. ICTP errors are 11 to 57$\times$ below the per-bus average of the context at full metering, and 1.3 to 4.0$\times$ below it with only 5% of measurements available. With measurements at six buses biased by 8$σ$, the ICTP angle error rises 1.8 to 5.0$\times$ over its bias-free value against 2.5 to 12.7$\times$ for WLS with bad-data detection. Without retraining, ICTP has lower voltage error than DSS$^2$ on both of its feeders, and lower than a retrained TAPINE on two of three reconfiguration sets.

eess.SY↗

Mechanistic Evidence for Spectral Structures in Prior-Data Fitted Networks

Prior-Data Fitted Networks (PFNs) perform approximate Bayesian inference in a single forward pass, and tabular foundation models (TFMs) built on them are now widely used. To understand what networks infer internally, recent mechanistic studies of TFMs locate where predictions form, but treat these models as tabular predictors rather than as PFNs. It therefore remains unknown whether PFNs represent the spectral content of their context, the quantity that specifies a stationary kernel, and whether this content can be read out as an explicit kernel. We answer both questions. First, across seven PFNs, including four pretrained TFMs and a model trained only on a decision-tree prior, a linear probe on the residual stream recovers the frequency of the context with $R^2 \geq 0.95$. This structure is led by a single principal direction. Second, activation and subspace patching show that the network uses the structure through a low-dimensional subspace, where a few spectral directions move predictions far more than random ones. This holds even for the decision-tree model, so a spectral training prior is not required. On real datasets with up to 499 features, 64 of the 192 directions of TabPFN, chosen without labels, carry 85 to 95\% of the causal effect of the context in all but one pair. Third, we introduce a Filter Bank Decoder that turns frozen PFN representations into an explicit stationary kernel through Bochner's theorem. Without any test-time optimization, the decoded kernel supports Gaussian process regression competitive with deep kernel learning and random Fourier features at about $200\times$ lower cost. PFN latents therefore hold spectral structure that is causally used and recoverable as a portable kernel.

cs.LG↗

Teaching LLMs to Generate Challenging MILP Instances via Solver Feedback

Generating optimization instances that are both feasible and computationally challenging is crucial for benchmarking solvers and training learning-based optimization algorithms. Existing non-LLM generators rely on seed instances or parameter tuning, resulting in high test-time computational cost, while existing LLM generators lack explicit hardness measures. Recent reinforcement learning methods with verifier feedback evaluate only binary correctness, which is misaligned with generating challenging problems. We note that an optimization solver reports the cost of solving at several stages of its pipeline, and leverage this to design a reward that scores both the solvability and the hardness of generated problems, measured by branch-and-bound nodes and post-cut relaxation gaps. Our key idea is a challenger-solver asymmetric self-play approach, where an LLM challenger generates progressively harder instances and the solver verifies feasibility and hardness, so no seed or training MILP instances are required. We fine-tune Gemma-4-12B and Qwen3.5-4B with GRPO and a size curriculum into OptiScribe-12B and OptiScribe-4B, which generate feasible yet challenging MILP problems from natural language instructions. On capacitated facility location and max-cut, OptiScribe-12B raises median SCIP search nodes by 1.7-5x and post-cut gaps by 1.1-1.7x over its base model and improves the feasibility rate on facility location by 9-19 points, while OptiScribe-4B raises median nodes by up to 15.6x. The problems cover a wider difficulty range than public benchmarks of the same size, follow instructions on density and difficulty, and can tune solver settings for families that public libraries lack. These results indicate that optimization-specific rewards, used in self-play mode, can teach LLMs to generate high-difficulty optimization benchmarks. We will release our code and models publicly on acceptance.

cs.AI↗

Coarse composition suffices: tabular in-context learning for multi-activity antimicrobial peptide profiling

Antimicrobial peptides (AMPs) often act against multiple pathogen classes, making multi-label activity prediction a more realistic screening target than binary antimicrobial classification. The ESCAPE benchmark formalizes this setting, but leading approaches typically rely on multimodal, structure-conditioned deep models that are costly to train and tune. We show that a simple, sequence-only pipeline can match and surpass these methods by combining 330 interpretable sequence descriptors with TabPFN, a tabular foundation model that performs in-context prediction in a single forward pass without gradient-based training or hyperparameter search. On ESCAPE (82,359 peptides; five labels), a label-powerset TabPFN model achieves mAP-5 = 77.8%, improving on the previously best reported 72.1%. A probabilistic classifier chain is the first method to match or exceed the best published average precision on each of the five labels simultaneously. The gains persist under the prior state-of-the-art single-fold training protocol, indicating they are not a training-set-size artefact, and are largest for remote homologues (+11.2 points below 30% sequence identity). Ablations further show that predicted structure is unnecessary at inference and that performance is not driven by any single descriptor family: ten global physicochemical scalars recover 91% of full-feature performance. Finally, explicitly modelling label dependence yields targeted benefits for scarce activities and supports ranking which activity to assay next from partial positive evidence.

cs.LG↗

Conditions for Quantum Advantage in AC Power Flow

This paper aims to contextualize the requirements for Quantum Computing (QC) algorithms to achieve a quantum advantage in solving the alternating current power flow (ACPF) problem, with a focus on runtime complexity. First, we establish a benchmark for a QC iterative solver to demonstrate an advantage over the classical Newton-Raphson Load Flow (NRLF) algorithm. Next, we derive a baseline expression for the end-to-end runtime complexity of any Gate-based QC algorithm as $Ω(N κ/\varepsilon),$ reflecting dependence on system size $N$, condition number $κ$, and error tolerance $\varepsilon$. Finally, we highlight key areas where QC algorithms may offer potential benefits over NRLF in addressing the standard ACPF problem.

quant-ph↗

Gradient-Free Topology Adaptation for Power Flow Surrogates via In-Context Whitening

Machine-learned surrogates for the AC power flow (ACPF) problem amortize the cost of repeated solves on a fixed network, but lose one to two orders of magnitude of accuracy when a line outage changes the topology. This degradation is an operator shift. The altered admittance matrix changes the input-to-output map, so identical inputs yield a different output distribution. Existing methods correct this with target-topology data and per-topology gradient steps. We ask whether the correction can instead be made statistical and gradient-free. We propose In-Context Whitening (ICW), which trains an ACPF surrogate in an output space whitened by the base topology's first two moments, and adapts it to an unseen N-1 or N-2 topology by re-estimating that whitening from a few hundred solved cases on the new topology. This adaptation is gradient-free, weight-free, and architecture-agnostic. We prove that among affine whiteners the unique choice that preserves the coordinate-wise semantics of the physical output vector is ZCA whitening, so within efficient invertible corrections, two moments are sufficient. Across the IEEE 30-, 118-, and 300-bus systems under N-1 and N-2 contingencies, ICW reduces overall error by 6$\times$ to 28$\times$ over frozen surrogates (up to 54$\times$ per-quantity under N-2) and cuts worst-bus power-balance mismatch by up to 30$\times$, with consistent gains across three backbones. At deployment scale it matches or beats gradient-based adaptation in accuracy while adapting 21$\times$ to 34$\times$ faster, with a cost that parallelizes on commodity CPU cores rather than requiring one GPU per contingency.

eess.SY↗

Towards AC Feasibility of DCOPF Dispatch

DC Optimal Power Flow (DCOPF) is widely utilized in power system operations due to its simplicity and computational efficiency. However, its lossless, reactive power-agnostic model often yields dispatches that are infeasible under practical operating scenarios such as the nonlinear AC power flow (ACPF) equations. While theoretical analysis demonstrates that DCOPF solutions are inherently AC-infeasible, their widespread industry adoption suggests substantial practical utility. This paper develops a unified DCOPF-ACPF pipeline to recover AC feasible solutions from DCOPF-based dispatches. The pipeline uses four DCOPF variants and applies AC feasibility recovery using both distributed slack allocation and PV/PQ switching. The main objective is to identify the most effective pipeline for restoring AC feasibility. Evaluation across over 10,000 dispatch scenarios on various test cases demonstrates that the structured ACPF model yields solutions that satisfy both the ACPF equations, and all engineering inequality constraints. In a 13,659 bus case, the mean absolute error and cost differences between DCOPF and ACOPF are reduced by 75% and 93%, respectively, compared to conventional single slack bus methods. Under extreme loading conditions, the pipeline reduces inequality constraint violations by a factor of 3 to 5.

eess.SY↗

PACR: Parameter-Optimized AC Power Flow Restoration for AC Feasible DCOPF Dispatch

The DC optimal power flow is widely used in power system operations because of its computational efficiency and scalability. However, DC dispatches are not guaranteed to satisfy the nonlinear AC power-flow equations or associated operational limits. This paper develops a parameterized, differentiable AC power-flow restoration method for mapping DC dispatches to AC-consistent operating points. The method incorporates distributed slack for active-power balancing and PV/PQ switching for reactive-power regulation, both implemented using smooth differentiable surrogates with tunable parameters, including slack participation factors, voltage setpoints, and regulation steepness. These parameters are trained offline by differentiating through the AC restoration equations using the implicit function theorem. Once trained, the optimized parameters are fixed and used directly during AC power-flow recovery from DC dispatches. The approach is evaluated on IEEE, ACTIVSg, and PEGASE test systems using setpoints computed by standard DC optimal power flow. Results show that the optimized restoration method improves AC feasibility recovery across various systems relative to conventional single-slack AC power-flow recovery. On the 9,241-bus case, the optimized method improves cost difference by 80% relative to the conventional recovery baseline and improves solving time relative to ACOPF by 75%.

eess.SY↗

Learning Power Flow with Confidence: A Probabilistic Guarantee Framework for Voltage Risk

The absence of formal performance guarantees in machine learning (ML) has limited its adoption for safety-critical power system applications, where confidence and interpretability are as vital as accuracy. In this work, we present a probabilistic guarantee for power flow learning and voltage risk estimation, derived through the framework of Gaussian Process (GP) regression. Specifically, we establish a bound on the expected estimation error that connects the GP's predictive variance to confidence in voltage risk estimates, ensuring statistical equivalence with Monte Carlo-based ACPF risk quantification. To enhance model learnability in the low-data regime, we first design the Vertex-Degree Kernel (VDK), a topology-aware additive kernel that decomposes voltage-load interactions into local neighborhoods for efficient large-scale learning. Building on this, we introduce a network-swipe active learning (AL) algorithm that adaptively samples informative operating points and provides a principled stopping criterion without requiring out-of-sample validation. Together, these developments mitigate the principal bottleneck of ML-based power flow, its lack of guaranteed reliability, by combining data efficiency with analytical assurance. Empirical evaluations across IEEE 118-, 500-, and 1354-bus systems confirm that the proposed VDK-GP achieves mean absolute voltage errors below 1E-03 p.u., reproduces Monte Carlo-level voltage risk estimates with 15x fewer ACPF computations, and achieves over 120x reduction in evaluation time while conservatively bounding violation probabilities.

eess.SY↗

Rethinking Neural Width for Alternating Current Optimal Power Flow Proxies

Deep learning proxies for Alternating Current Optimal Power Flow (ACOPF) lack systematic methods for determining architectural size. This paper conducts a constructive thought experiment to answer a fundamental inquiry: how wide must a neural network be to almost accurately approximate the ACOPF manifold? We introduce a Loss-Guided Neural Densification (LG-ND) algorithm that incrementally discovers necessary capacity by expanding only when the current deep neural network topology fails to improve further. Empirical results across various IEEE systems show that LG-ND achieves performance parity with literature baselines using up to ten times fewer neurons per layer. Such architectural minimalism is critical for the formal verification required in safety-critical grid operations.

cs.LG↗

Structural Disentanglement in Bilinear MLPs via Architectural Inductive Bias

Selective unlearning and long-horizon extrapolation remain fragile in modern neural networks, even when tasks have underlying algebraic structure. In this work, we argue that these failures arise not solely from optimization or unlearning algorithms, but from how models structure their internal representations during training. We explore if having explicit multiplicative interactions as an architectural inductive bias helps in structural disentanglement, through Bilinear MLPs. We show analytically that bilinear parameterizations possess a `non-mixing' property under gradient flow conditions, where functional components separate into orthogonal subspace representations. This provides a mathematical foundation for surgical model modification. We validate this hypothesis through a series of controlled experiments spanning modular arithmetic, cyclic reasoning, Lie group dynamics, and targeted unlearning benchmarks. Unlike pointwise nonlinear networks, multiplicative architectures are able to recover true operators aligned with the underlying algebraic structure. Our results suggest that model editability and generalization are constrained by representational structure, and that architectural inductive bias plays a central role in enabling reliable unlearning.

cs.LG↗

Fast Diffusion with Physics-Correction for ACOPF

Generating large-scale, physically consistent AC Optimal Power Flow (ACOPF) datasets is essential for modern data-driven power system applications. The central challenge lies in balancing solution accuracy with computational efficiency. Recent diffusion-based generative models produce high-quality samples; however, their slow sampling procedures limit practical scalability. In this work, we argue that exact physical feasibility is ultimately enforced by power flow solvers or projection steps, and therefore the generative model only needs to produce good initializations rather than perfectly feasible solutions. Based on this insight, we propose a fast diffusion framework using Denoising Diffusion Implicit Models (DDIM) combined with physics-guided corrections during sampling. The proposed method replaces slow stochastic refinement with a small number of deterministic steps and explicit constraint guidance. Experiments on IEEE 6-, 24-, and 118-bus systems show that our approach achieves up to 20 times faster sampling than standard diffusion models while maintaining comparable statistical accuracy and physical consistency. This makes the method well suited for scalable OPF dataset generation and practical power system learning tasks. We release the implementation code at https://github.com/PSquare-Lab/DDIM_OPF.

eess.SY↗

Decoupled-Value Attention for Prior-Data Fitted Networks: GP Inference for Physical Equations

Prior-data fitted networks (PFNs) are a promising alternative to time-consuming Gaussian process (GP) inference for creating fast surrogates of physical systems. PFN reduces the computational burden of GP-training by replacing Bayesian inference in GP with a single forward pass of a learned prediction model. However, with standard Transformer attention, PFNs show limited effectiveness on high-dimensional regression tasks. We introduce Decoupled-Value Attention (DVA)-- motivated by the GP property that the function space is fully characterized by the kernel over inputs and the predictive mean is a weighted sum of training targets. DVA computes similarities from inputs only and propagates labels solely through values. Thus, the proposed DVA mirrors the GP update while remaining kernel-free. We demonstrate that PFNs are backbone architecture invariant and the crucial factor for scaling PFNs is the attention rule rather than the architecture itself. Specifically, our results demonstrate that (a) localized attention consistently reduces out-of-sample validation loss in PFNs across different dimensional settings, with validation loss reduced by more than 50% in five- and ten-dimensional cases, and (b) the role of attention is more decisive than the choice of backbone architecture, showing that CNN, RNN and LSTM-based PFNs can perform at par with their Transformer-based counterparts. The proposed PFNs provide 64-dimensional power flow equation approximations with a mean absolute error of the order of E-03, while being over 80x faster than exact GP inference.

cs.LG↗

Small LLMs with Expert Blocks Are Good Enough for Hyperparamter Tuning

Hyper-parameter Tuning (HPT) is a necessary step in machine learning (ML) pipelines but becomes computationally expensive and opaque with larger models. Recently, Large Language Models (LLMs) have been explored for HPT, yet most rely on models exceeding 100 billion parameters. We propose an Expert Block Framework for HPT using Small LLMs. At its core is the Trajectory Context Summarizer (TCS), a deterministic block that transforms raw training trajectories into structured context, enabling small LLMs to analyze optimization progress with reliability comparable to larger models. Using two locally-run LLMs (phi4:reasoning14B and qwen2.5-coder:32B) and a 10-trial budget, our TCS-enabled HPT pipeline achieves average performance within ~0.9 percentage points of GPT-4 across six diverse tasks.

cs.LG↗

Limitations of Fault-Tolerant Quantum Linear System Solvers for Quantum Power Flow

Quantum computers hold promise for solving problems intractable for classical computers, especially those with high time or space complexity. Practical quantum advantage can be said to exist for such problems when the end-to-end time for solving such a problem using a classical algorithm exceeds that required by a quantum algorithm. Reducing the power flow (PF) problem into a linear system of equations allows for the formulation of quantum PF (QPF) algorithms, which are based on solving methods for quantum linear systems such as the Harrow-Hassidim-Lloyd (HHL) algorithm. Speedup from using QPF algorithms is often claimed to be exponential when compared to classical PF solved by state-of-the-art algorithms. We investigate the potential for practical quantum advantage in solving QPF compared to classical methods on gate-based quantum computers. Notably, this paper does not present a new QPF solving algorithm but scrutinizes the end-to-end complexity of the QPF approach, providing a nuanced evaluation of the purported quantum speedup in this problem. Our analysis establishes a best-case bound for the HHL-based quantum power flow complexity, conclusively demonstrating that the HHL-based method has higher runtime complexity compared to the classical algorithm for solving the direct current power flow (DCPF) and fast decoupled load flow (FDLF) problem. Notably, our analysis and conclusions can be extended to any quantum linear system solver with rigorous performance guarantees, based on the known complexity lower bounds for this problem. Additionally, we establish that for potential practical quantum advantage (PQA) to exist it is necessary to consider DCPF-type problems with a very narrow range of condition number values and readout requirements.

quant-ph↗

Optimization Proxies using Limited Labeled Data and Training Time -- A Semi-Supervised Bayesian Neural Network Approach

Constrained optimization problems arise in various engineering systems such as inventory management and power grids. Standard deep neural network (DNN) based machine learning proxies are ineffective in practical settings where labeled data is scarce and training times are limited. We propose a semi-supervised Bayesian Neural Networks (BNNs) based optimization proxy for this complex regime, wherein training commences in a sandwiched fashion, alternating between a supervised learning step for minimizing cost, and an unsupervised learning step for enforcing constraint feasibility. We show that the proposed semi-supervised BNN outperforms DNN architectures on important non-convex constrained optimization problems from energy network operations, achieving up to a tenfold reduction in expected maximum equality gap and halving the inequality gaps. Further, the BNN's ability to provide posterior samples is leveraged to construct practically meaningful probabilistic confidence bounds on performance using a limited validation data, unlike prior methods. The implementation code for this study is available at: https://github.com/kaarthiksundar/BNN-OPF/.

cs.LG↗

Gaussian Processes in Power Systems: Techniques, Applications, and Future Works

The increasing integration of renewable energy sources (RESs) and distributed energy resources (DERs) has significantly heightened operational complexity and uncertainty in modern power systems. Concurrently, the widespread deployment of smart meters, phasor measurement units (PMUs) and other sensors has generated vast spatiotemporal data streams, enabling advanced data-driven analytics and decision-making in grid operations. In this context, Gaussian processes (GPs) have emerged as a powerful probabilistic framework, offering uncertainty quantification, non-parametric modeling, and predictive capabilities to enhance power system analysis and control. This paper presents a comprehensive review of GP techniques and their applications in power system operation and control. GP applications are reviewed across three key domains: GP-based modeling, risk assessment, and optimization and control. These areas serve as representative examples of how GP can be utilized in power systems. Furthermore, critical challenges in GP applications are discussed, and potential research directions are outlined to facilitate future power system operations.

eess.SY↗