Searcharxiv⌕ Search

arXiv subjects

Feng Ji

Publications and source records attributed to Feng Ji.

At least 19 recordsLinked to original sources
arXiv2026

Versatile Graph Foundation Models from Frozen Time-Series Foundation Models via Parallel Flows

Graph foundation models aim to serve unseen graphs across tasks, feature spaces, and label sets. Most current models learn how to represent a new graph by pre-training on other graphs. Because graph data are scarce and differ across domains in features and structures, what they learn may not transfer to a graph from an unseen domain. We show that a frozen time-series foundation model, pre-trained on abundant time series, can instead serve as a graph foundation model without graph pre-training and without any gradient update on the target graph. Such a model reads ordered multivariate sequences, but a graph has no time axis, and covering all its edges with sequences is nontrivial because a node can have many neighbors. We therefore propose GraphFlowFM, which decomposes a graph into parallel flows, sets of vertex-disjoint paths that jointly cover every edge, and reads each path as a multivariate time series with the nodes as time steps. A frozen TimesFM encodes the paths with causal attention, which gives each node a self token and a context token, and a closed-form ridge readout weighs these tokens by their alignment with the labels. In node classification, GraphFlowFM attains the best average rank on 21 datasets among graph foundation models published in 2025 and 2026, ranks first on 17 datasets, and leads on six of the seven heterophilous datasets. It also exceeds pre-trained graph foundation models in link prediction on all six datasets and ranks second in molecule classification, behind only a model fine-tuned after pre-training on molecules.

cs.LG↗
arXiv2026

PolarFlow: Spectral Geometric Flow Matching for Directed Graph Generation via Polar Decomposition

Directed graphs model asymmetric relationships in biology, transportation, social networks, and electronic design automation. Recent discrete diffusion and flow-matching methods generate directed graphs effectively, but their edge-wise categorical formulations require dense predictions over all ordered node pairs and repeated categorical updates during generation. This motivates a fundamental question: can comparable generative quality be achieved through a more computationally efficient formulation that better exploits the structure of directed adjacency matrices? Existing benchmarks also cover limited network structures, restricting comprehensive evaluation. We introduce PolarFlow, a continuous geometric flow-matching framework based on polar decomposition and truncated spectral coordinates. Its representation comprises four coupled components: a Stiefel frame and rotation angles encoding the orthogonal factor, and a second Stiefel frame and nonnegative eigenvalues characterizing the positive-semidefinite factor. PolarFlow learns a joint velocity field along geometry-aware conditional paths on this product space, reconstructs a continuous adjacency-score matrix, and discretizes it into a binary directed graph. By replacing discrete edge-wise generation with joint continuous transport, PolarFlow improves computational efficiency while retaining global adjacency structure. To broaden evaluation, we introduce complementary Circuit and Twitter benchmarks spanning engineered functional and naturally evolved social networks. Across synthetic and real-world datasets, PolarFlow achieves generative quality comparable to competitive discrete methods while substantially reducing training and sampling time.

eess.SP↗
arXiv2026

Timbre: Reliability-aware harmonic consensus for label-free multimodal test-time rexinforcement learning

Label-free test-time reinforcement learning adapts vision--language models using rewards derived from their own predictions. However, misinterpreted visual evidence and sensitivity to textual formulation can produce a consistent but incorrect consensus that majority voting reinforces. Cross-view aggregation can expose such errors, but image and text transformations may themselves degrade task-relevant evidence to different extents. We introduce \method, a reliability-aware harmonic consensus framework for multimodal test-time reinforcement learning. \method constructs a $2\times2$ grid combining original and perturbed images with direct-solving and reframing modes, selects pseudo-labels through weighted harmonic aggregation, and adapts modality-level weights using batch-averaged, temporally smoothed shifts in pseudo-label confidence as a heuristic reliability signal. We formulate consensus as a weighted $χ^2$-barycenter problem whose unique minimizer is the normalized weighted harmonic aggregate, without requiring correct-answer invariance. An exact support--dispersion decomposition characterizes when harmonic aggregation preserves or corrects arithmetic rankings, while a sensitivity analysis establishes how modality reweighting changes pairwise ranking margins, informing interpretation of the heuristic updates. Across six benchmarks and three Qwen3-VL model scales, \method outperforms TTRV in 16 of 18 benchmark--model settings on adaptation instances and 17 of 18 settings on held-out instances under matched estimation and optimization rollout counts. These results demonstrate improved target-task adaptation and transfer to unseen instances without further updates.

cs.AI↗
arXiv2026

Graph Spectral Series and Its Signatures

Bandlimitedness is a central modeling assumption in graph signal processing (GSP), but many graph signals exhibit substantial energy throughout the graph-frequency domain. We introduce a new framework for analyzing such non-bandlimited signals by treating their graph Fourier spectra as ordered sequences and representing them with path signatures. Specifically, motivated by the observation that non-bandlimited graph spectra often display rough, time-series like behavior, we introduce the notion of graph spectral series and develop a signature-based representation that captures global spectral structure without frequency truncation. We establish conditions under which the proposed representation is continuous and characterize its information-preserving properties. We further provide an asymptotic interpretation of graph spectral series for sequences of expanding graphs. Experiments on real-world trend-prediction and change-detection tasks show that signature features can improve performance over representations based directly on the original graph signals. The proposed approach provides a novel principled signature-based tool for analyzing non-bandlimited graph signals.

eess.SP↗
arXiv2026

Constructive Port-Hamiltonian Energy Shaping Design of Dispatchable Virtual Oscillators in Grid-Forming Converters

Port-Hamiltonian (PH) theory offers a passivity-based framework for grid-forming control, yet conventional dispatchable virtual oscillator control (dVOC) does not naturally admit a dissipative PH realization, since its amplitude egulation, synchronization, and power dispatch are inherently coupled without a unified energy interpretation. This paper formulates the outer-loop dynamics as a dissipative PH system, thereby unifying amplitude regulation, synchronization, and power dispatch within one energy structure. The formulation rests on the key property that logistic-type radial regulation, characterized by an inherent saturation-like nonlinearity, permits an exact gradient decomposition compatible with the quadratic energy storage. On this basis, a unified shaped Hamiltonian is constructed, which encapsulates both amplitude restoration and power dispatch. Radial gain matching derived from this Hamiltonian yields explicit closed-form arameter inequalities that guarantee almost-global asymptotic stability and local exponential convergence. Moreover, tuning the power-error weighting coefficient actively shapes the energy landscape, thereby eliminating the undesirable low-voltage power-flow solution from the stationary set and ensuring convergence to the desired high-voltage equilibrium point. The Hessian singularity condition further provides the critical weight threshold that guarantees uniqueness of the high-voltage equilibrium. Numerical simulations validate the proposed method.

eess.SY↗
arXiv2026

From Rigid to Adiabatic: Canonical Regularization of AC Networks via Action-Angle Variables

Traditional power system analysis relies on timescale separation and the rigid-network assumption, freezing electromagnetic transients into algebraic power-flow equations via Steinmetz's phasor theory. As grid-forming converter penetration increases, magnetic energy dynamics on transmission lines interact with converter control loops on comparable timescales, challenging this rigid-network assumption. Returning to Faraday's law of electromagnetic induction driven by rotating magnetic fields, this paper models the transmission lines' rotating magnetic fields in action-angle canonical coordinates, regularizes the rigid algebraic constraints of power-flow equations into canonical equations on adiabatic symplectic manifolds, and establishes a port-Hamiltonian standard form for AC power grids. Based on the minimal-counterexample principle and using the equal-area criterion's classical two-machine system, this paper reveals a latitudinal instability channel via Bloch-sphere coordinates: Q-V control releases voltage-amplitude freedom, shifting the stability boundary from the equatorial UEP (unstable equilibrium point) to a saddle point, thereby unifying the analytical frameworks of P-delta angle stability and Q-V voltage stability in power system analysis.

eess.SY↗
arXiv2026

LLM-generated personalized nudges for improving pro-environmental behavior: Field evidence from resource conservation

Encouraging pro-environmental behavior remains a major challenge for sustainable cities. Conventional feedback nudges can show individuals how their current behavior compares with environmental goals but often provide limited guidance on what to do differently in daily life. This study examines whether supplementing weekly feedback on participants' behavior with LLM-generated personalized action suggestions improves pro-environmental behavior, using daily electricity and hot-water conservation as a case study. We developed an LLM agent that generated weekly conservation messages from participant profiles, recent consumption records, and prior interaction history, combining a usage report with personalized suggestions, behavioral-change scenarios, and estimated savings. The agent was evaluated in a three-arm randomized field experiment with 233 university residents in Beijing from November 2024 to January 2025. Participants received text-based nudges, image-enhanced nudges, or LLM-generated personalized nudges over five intervention rounds. Daily electricity use and shower hot-water use were measured using dormitory meter readings and billing records. Compared with text-based feedback, LLM-generated personalized nudges reduced electricity consumption by 0.56 kWh per room-day (p = 0.014), corresponding to an 18.3 percentage-point higher saving rate. Image-enhanced feedback alone showed no clear improvement. Hot-water savings followed the same direction but were smaller and less precisely estimated (9.8 percentage points, p = 0.087). Personalized nudges contained more planning, appliance-specific, and action-oriented language and were associated with more sustained, task-focused engagement. These findings offer a pathway for integrating generative AI into sustainable urban management.

cs.CY↗
arXiv2026

Statistical realism is not evidence that LLMs can estimate treatment effects in social science experiments

Large language models (LLMs) are increasingly used to simulate human responses and estimate treatment effect of interventions when real-world experiments are costly or infeasible. The treatment-effect estimates are often evaluated using statistical realism, the degree to which simulated responses reproduce properties of observed human responses, although whether realism predicts treatment-effect accuracy remains unknown. Here we test this proxy relationship by jointly measuring statistical realism and treatment-effect accuracy on the same simulated responses in a cross-national experiment with 59,508 participants from 62 countries using three LLMs. The correlation between statistical realism and treatment-effect accuracy is weak, and optimizing for statistical realism can even worsen treatment-effect accuracy when selecting models, prompts, and target populations. The pattern replicates in two additional cross-national experiments spanning 12 and 27 countries with 20,785 participants. The divergence between the two reflects distinct error structures and is larger for behavioral outcomes, where models appear to extrapolate behavioral effects from attitudinal patterns. Because this divergence may remain hidden in deployment, errors can propagate into simulation-informed decisions. We introduce a diagnostic framework for LLM-generated synthetic data and discuss how treatment-effect validation should proceed under varying availability of experimental benchmarks. Simulated responses and simulated treatment effects are distinct estimation targets, and evidence for one does not certify the other.

cs.CY↗
arXiv2026

Filter Learning for Subgraphs: Algebras and Performance Risk Bounds

Graph signal processing tasks that leverage spectral information typically assume access to the complete graph topology, which is often unavailable in practice. We propose a systematic framework for subgraph filter learning (SFL), where subgraph-supported operators approximate ambient graph filters under partial observations. We formulate SFL as a statistical learning problem in which optimal subgraph operators are inherently data-dependent. To address the difficulty of directly estimating such operators, we develop a subgraph filter algebra based on distance-aware Laplacian constructions, defining a structured and controllable class of filters for effective approximation. We further establish performance risk bounds under the least squares loss, quantifying how well the learned operator approximates the restricted ambient mapping. Experiments real-world datasets show that, for SFL tasks, the proposed algebraic models consistently outperform polynomial filters, distribution-agnostic operators, and direct numerical filter learning baselines that attempt to recover the underlying structure from data.

cs.LG↗
arXiv2026

Optimal Sensor Placement via Graph-constrained Flow Matching

Optimal sensor placement is a fundamental problem in graph signal processing (GSP), where a limited number of sensors are deployed to reconstruct a continuous signal field. Existing GSP methods rely on combinatorial optimization over discretized graphs, resulting in high computational cost and sensor locations restricted to graph vertices. We reformulate sensor placement as a continuous-space generative modeling problem. An offline GSP optimization routine first generates training samples of optimal sensor configurations, from which a flow matching (FM) model learns their distribution. At inference, the learned velocity field directly generates continuous sensor coordinates, eliminating online combinatorial optimization. We further develop a permutation-invariant conditional generation framework for deployment with fixed anchor sensors. Experiments on a realistic radio propagation simulator demonstrate the effectiveness of the proposed framework for sensor placement.

eess.SP↗
arXiv2026

Graph Distribution-valued Signals in Wasserstein Spaces: Theory and Applications

We introduce a framework for graph signal processing (GSP) in which signals are represented as graph distribution-valued signals (GDSs), i.e., probability measures in a Wasserstein space. This perspective addresses fundamental limitations of classical vector-based GSP, including the requirement for complete synchronous observations across vertices and the need for strict temporal correspondence in observed filter input--output pairs. Furthermore, by modeling the graph structure as a distribution conditioned on signal realizations, we provide a principled approach to signal-dependent graph structures, which are common in real-world applications, while explicitly encoding uncertainty in graph topology. Our framework inherently captures uncertainty and stochasticity while strictly generalizing traditional graph signals, which can be interpreted as Dirac delta measures. We develop a systematic correspondence between foundational GSP concepts and their GDS analogs, showing that classical formulations emerge as special cases of our framework. We establish theoretical continuity results for GDS transforms, providing stability guarantees for input perturbations and distribution approximations. We demonstrate the utility of this approach through example applications, including graph filter learning and anomaly detection, and validate its effectiveness through empirical studies.

eess.SP↗
arXiv2026

Port-Hamiltonian Control and Structure-Preserving Algorithm for Grid-Forming SVGs

This paper presents a port-Hamiltonian (PH) modeling, control, and structure-preserving simulation framework for grid-forming static var generators (SVGs). A PH model is established that captures energy exchange among the inductor, capacitor, and DC-link storage ports. Since external disturbances cannot be fully canceled by feedback, an input-to-state stable (ISS) controller is designed to steer subsystem states to zero while minimizing disturbance effects. The controller contains only three tunable parameters with clear physical interpretations and is robust against input errors. A Dirac-structure-preserving midpoint rule is developed, which exactly conserves the Hamiltonian energy when disturbances are absent. Numerical comparisons show that the ISS controller achieves faster settling, smaller offset, and lower control effort than a conventional PI controller, and the structure-preserving midpoint rule maintains exact energy conservation and superior long-term accuracy over standard Runge-Kutta methods.

math.OC↗
arXiv2026

Less is More: Towards Simple Graph Contrastive Learning

Graph Contrastive Learning (GCL) has shown strong promise for unsupervised graph representation learning, yet its effectiveness on heterophilic graphs, where connected nodes often belong to different classes, remains limited. Most existing methods rely on complex augmentation schemes, intricate encoders, or negative sampling, which raises the question of whether such complexity is truly necessary in this challenging setting. In this work, we revisit the foundations of supervised and unsupervised learning on graphs and uncover a simple yet effective principle for GCL: mitigating node feature noise by aggregating it with structural features derived from the graph topology. This observation suggests that the original node features and the graph structure naturally provide two complementary views for contrastive learning. Building on this insight, we propose an embarrassingly simple GCL model that uses a GCN encoder to capture structural features and an MLP encoder to isolate node feature noise. Our design requires neither data augmentation nor negative sampling, yet achieves state-of-the-art results on heterophilic benchmarks with minimal computational and memory overhead, while also offering advantages in homophilic graphs in terms of complexity, scalability, and robustness. We provide theoretical justification for our approach and validate its effectiveness through extensive experiments, including robustness evaluations against both black-box and white-box adversarial attacks.

cs.LG↗
arXiv2026

Adaptive Multi-view Graph Contrastive Learning via Fractional-order Neural Diffusion Networks

Graph contrastive learning (GCL) learns node and graph representations by contrasting multiple views of the same graph. Existing methods typically rely on fixed, handcrafted views-usually a local and a global perspective, which limits their ability to capture multi-scale structural patterns. We present an augmentation-free, multi-view GCL framework grounded in fractional-order continuous dynamics. By varying the fractional derivative order $α\in (0,1]$, our encoders produce a continuous spectrum of views: small $α$ yields localized features, while large $α$ induces broader, global aggregation. We treat $α$ as a learnable parameter so the model can adapt diffusion scales to the data and automatically discover informative views. This principled approach generates diverse, complementary representations without manual augmentations. Extensive experiments on standard benchmarks demonstrate that our method produces more robust and expressive embeddings and outperforms state-of-the-art GCL baselines.

cs.LG↗
arXiv2026

Beyond Accuracy: Evaluating Strategy Diversity in LLM Mathematical Reasoning

Large language models now achieve high final-answer accuracy on mathematical reasoning benchmarks, but accuracy alone does not capture reasoning flexibility. We introduce a strategy-level evaluation framework instantiated on 80 AMC 10/12 and AIME problems with 217 AoPS-derived reference strategy families. Model outputs are annotated for strategy identity, validity, and correctness using dual-AI coding with human adjudication. Across four frontier models, we find a pronounced decoupling between answer accuracy and strategy diversity. Under a single-solution prompt, all models achieve high accuracy (95%-100%), but under a multiple-strategy prompt they recover substantially fewer strategies than the human reference set. Gemini, DeepSeek, GPT, and Claude generate 184, 152, 151, and 110 distinct valid strategies, respectively, with the largest gaps in Geometry and Number Theory. The models collectively produce 50 benchmark-novel valid strategies, indicating both incomplete coverage of human strategies and some capacity for alternative reasoning. A repeated-run robustness check on 20 problems shows diminishing gains in discovered strategies, with the strongest model recovering only 39 of 55 AoPS-reference strategies (71%) after three runs. These findings position strategy diversity as a complementary dimension for evaluating mathematical reasoning beyond answer correctness.

cs.AI↗
arXiv2026

Reinforcement Learning Measurement Model

Interactive assessments generate sequential process data that are not well handled by conventional item response models. Existing MDP-based measurement approaches, such as the Markov decision process measurement model (MDP-MM, LaMar, 2018), link action choices to state-action values, but their reliance on person-specific tabular value functions makes them difficult to scale beyond small, fully enumerated tasks. We propose the Reinforcement Learning Measurement Model (RLMM), a measurement framework that decouples person-level choice sensitivity from task-level value representation through a shared parametric action-value function, making estimation more computationally efficient for larger process-data settings. The model combines a Boltzmann choice rule with normalized advantages, a soft Bellman consistency penalty, and a block-coordinate MAP procedure for joint estimation, while also yielding step-level influence diagnostics for identifying behaviorally critical decisions. In peg-solitaire simulations, the RLMM achieved higher estimation accuracy and substantially lower runtime than the original MDP-MM, with advantages increasing as task complexity grew. In AQUALAB gameplay logs, the estimated person parameter was positively associated with cumulative reward, task completion, and behavioral efficiency. These results show that the RLMM extends decision-process-based psychometric models to larger and more behaviorally realistic environments while preserving an interpretable latent trait tied to decision making steps.

stat.ME↗
arXiv2026

Federated Item Response Models: A Gradient-driven Privacy-preserving Framework for Distributed Psychometric Estimation

Item Response Theory (IRT) models are widely used to estimate respondents' latent abilities and calibrate item difficulty. Traditional IRT estimation typically requires centralizing all raw responses, raising privacy and governance concerns. We introduce Federated Item Response Theory (FedIRT), a framework that enables distributed calibration of standard IRT models without transferring individual-level data, thereby preserving confidentiality while retaining statistical efficiency. To provide formal protection, we further develop FedIRT-DP, a user-level differentially private extension. Each site computes per-student gradients, clips them to a fixed norm, and shares only masked sums; the server adds calibrated Gaussian noise and performs MAP updates. This yields an auditable $(\varepsilon,δ)$ guarantee at the student level and a single, tunable privacy-utility trade-off via the clipping bound and noise scale. The same mechanism improves robustness to extreme response rows (e.g., all-zeros/ones). Across simulations, FedIRT matches the accuracy of centralized estimators from popular $\texttt{R}$ packages while avoiding data pooling; FedIRT-DP achieves comparable accuracy under stronger privacy and exhibits superior robustness to contamination. An empirical study on a real exam dataset demonstrates practical viability and consistent item and site-effect estimates. To facilitate adoption, we release an open-source $\texttt{R}$ package, $\texttt{FedIRT}$, implementing the two-parameter logistic (2PL) and partial credit models (PCM) with federated and differentially private training.

cs.LG↗
arXiv2026

Robust Standard Errors for Bayesian Posterior Functionals via the Infinitesimal Jackknife

Quantitative research in the social and behavioral sciences relies heavily on nonlinear posterior functionals such as indirect effects, standardized coefficients, effect sizes, intraclass correlations, and multilevel variance-explained measures. The posterior standard deviation (PostSD) is the default uncertainty summary for these quantities, yet it presupposes a correctly specified model. When the working model is wrong, as is common with behavioral data that exhibit heavy tails and heteroskedasticity, PostSD can severely underestimate the frequentist standard error. The nonparametric bootstrap offers robustness but requires repeated MCMC refits, while the delta method demands a separate analytic gradient derivation for every new functional. The infinitesimal jackknife standard error (Giordano & Broderick, 2023) sidesteps both limitations: it approximates the bootstrap variance through influence functions computed from a single MCMC run, applies to any posterior functional without modification, and requires no analytic derivatives. We discuss the use the IJSE methodology at both the observation level and the cluster level and evaluate it through four simulation studies covering six functionals from mediation analysis, ANOVA, and multilevel modeling, which are commonly used in the social and behavioral sciences. Under misspecification, PostSD substantially underestimated the true standard error across all settings, whereas IJSE closely tracked the bootstrap at a fraction of the computational cost. Under correct specification all three methods agreed, confirming that IJSE introduces no distortion when the model is right. These results show IJSE as a practical, general-purpose tool for robust uncertainty quantification in Bayesian workflows throughout the social and behavioral sciences

stat.ME↗