arXiv subjects

Yinsong Chen

Publications and source records attributed to Yinsong Chen.

10 recordsLinked to original sources

arXiv2026

Post-Hoc Uncertainty-Aware Explanations for Deployed Power Quality Disturbance Classifiers via Laplace Approximation

Deep learning classifiers achieve high accuracy in power quality disturbance (PQD) recognition, but existing explanation methods return a single deterministic attribution map and provide no measure of its reliability. This paper develops a post-hoc Bayesian explanation (B-explanation) method for trained PQD classifiers. A computationally efficient Laplace approximation converts the trained network into an approximate parameter posterior without retraining, and occlusion sensitivity is propagated through posterior samples to produce a distribution over disturbance-localization maps. Percentile summaries of this distribution yield explanations with distribution-free coverage bands: consensus summaries at low percentiles sharpen localization significantly for distinctive events such as sags, swells, and oscillatory transients, the band width indicates the reliability of each attribution, and the remaining disturbance types show class-dependent behavior. Explanation dispersion also increases under injected measurement noise and synthetic-to-field transfer, complementing predictive uncertainty. Experiments on a synthetic benchmark of 15 disturbance classes and on field-recorded sags compare the method with Monte Carlo dropout and deep ensembles under a common evaluation protocol, evaluate it against deterministic occlusion, LIME, and SHAP with localization and faithfulness metrics, and characterize the computational cost of explanation generation for grid monitoring applications.

cs.LG↗

arXiv2026

A Unified Framework for Uncertainty-Aware Explainable Artificial Intelligence: A Case Study in Power Quality Disturbance Classification

Post-hoc explainable AI (XAI) methods usually return one attribution map, even when the model represents uncertainty in its parameters. We define the \emph{explanation distribution} as the distribution of attribution maps obtained from sampled models. The uncertainty-aware relevance attribution operator (UA-RAO) summarises this distribution using the mean, dispersion, quantiles, and agreement sets. The theory separates posterior-approximation error from finite-sample error and accounts for changes across activation boundaries and for stochastic explainers. On a 15-class power-quality-disturbance benchmark, the mean occlusion explanation from a deep ensemble aligns better with known disturbance regions than the deterministic baseline, although the improvement depends on the disturbance type. Tests with controlled input distortions show that additive noise changes the explanations more than amplitude scaling or aligned temporal shifts.

cs.LG↗

arXiv2026

Uncertainty-Disentangled Probabilistic Stability Analysis in Wind Power Integrated Weak Grids

Conventional probabilistic small-signal stability analysis (PSSSA) propagates a single forecast distribution, conflating irreducible weather randomness (aleatoric) with reducible forecast-model uncertainty (epistemic). This letter propagates a second-order renewable forecast through the modal-stability map via an independent \emph{germ} variable, separating the two contributions exactly in closed form by a disentangled polynomial chaos expansion (d-PCE). The split underpins a forecast-aware $(α,β,γ)$ stability certificate whose conservative branch converges to its irreducible aleatoric limit at $O(N^{-1/2})$ -- making a failed certificate diagnostic: epistemic-dominated risk recovers with better data; aleatoric-dominated risk needs improvements of the physical control system.

eess.SY↗

arXiv2026

A Posterior-Predictive Variance Decomposition for Epistemic and Aleatoric Uncertainty in Wind Power Forecasting

Accurate wind power forecasting requires reliable uncertainty quantification, yet most existing methods report a single predictive uncertainty that conflates epistemic and aleatoric sources. This paper applies the law of total variance to the joint setting of heteroscedastic neural network regression and Bayesian posterior approximation, deriving an explicit decomposition of total uncertainty (TU) into aleatoric (AU) and epistemic (EU) components. The resulting estimators are compatible with standard posterior-approximation methods and with $β$-NLL training to regulate the mean--variance learning trade-off. A wind power--specific evaluation framework is proposed to validate disentanglement without access to ground-truth uncertainty labels, comprising three modules: controlled synthetic experiments to verify responses to heteroscedastic noise and distribution shift; data-property--driven validation on a real-world wind turbine SCADA dataset; and dataset-size scaling experiments to examine the predicted asymptotic behavior of EU. Across synthetic and real-world experiments, the decomposed AU and EU components respond in theoretically consistent directions to noise structure, distributional shift, and training-scale variation, supporting the theoretical consistency and operational utility of the proposed decomposition and evaluation protocol.

cs.LG↗

arXiv2025

Addressing the Inconsistency in Bayesian Deep Learning via Generalized Laplace Approximation

In recent years, inconsistency in Bayesian deep learning has attracted significant attention. Tempered or generalized posterior distributions are frequently employed as direct and effective solutions. Nonetheless, the underlying mechanisms and the effectiveness of generalized posteriors remain active research topics. In this work, we interpret posterior tempering as a correction for model misspecification via adjustments to the joint probability, and as a recalibration of priors by reducing aleatoric uncertainty. We also introduce the generalized Laplace approximation, which requires only a simple modification to the Hessian calculation of the regularized loss and provides a flexible and scalable framework for high-quality posterior inference. We evaluate the proposed method on state-of-the-art neural networks and real-world datasets, demonstrating that the generalized Laplace approximation enhances predictive performance.

cs.LG↗

arXiv2025

Red-Team Multi-Agent Reinforcement Learning for Emergency Braking Scenario

Current research on decision-making in safety-critical scenarios often relies on inefficient data-driven scenario generation or specific modeling approaches, which fail to capture corner cases in real-world contexts. To address this issue, we propose a Red-Team Multi-Agent Reinforcement Learning framework, where background vehicles with interference capabilities are treated as red-team agents. Through active interference and exploration, red-team vehicles can uncover corner cases outside the data distribution. The framework uses a Constraint Graph Representation Markov Decision Process, ensuring that red-team vehicles comply with safety rules while continuously disrupting the autonomous vehicles (AVs). A policy threat zone model is constructed to quantify the threat posed by red-team vehicles to AVs, inducing more extreme actions to increase the danger level of the scenario. Experimental results show that the proposed framework significantly impacts AVs decision-making safety and generates various corner cases. This method also offers a novel direction for research in safety-critical scenarios.

cs.LG↗

arXiv2025

Dynamic Residual Safe Reinforcement Learning for Multi-Agent Safety-Critical Scenarios Decision-Making

In multi-agent safety-critical scenarios, traditional autonomous driving frameworks face significant challenges in balancing safety constraints and task performance. These frameworks struggle to quantify dynamic interaction risks in real-time and depend heavily on manual rules, resulting in low computational efficiency and conservative strategies. To address these limitations, we propose a Dynamic Residual Safe Reinforcement Learning (DRS-RL) framework grounded in a safety-enhanced networked Markov decision process. It's the first time that the weak-to-strong theory is introduced into multi-agent decision-making, enabling lightweight dynamic calibration of safety boundaries via a weak-to-strong safety correction paradigm. Based on the multi-agent dynamic conflict zone model, our framework accurately captures spatiotemporal coupling risks among heterogeneous traffic participants and surpasses the static constraints of conventional geometric rules. Moreover, a risk-aware prioritized experience replay mechanism mitigates data distribution bias by mapping risk to sampling probability. Experimental results reveal that the proposed method significantly outperforms traditional RL algorithms in safety, efficiency, and comfort. Specifically, it reduces the collision rate by up to 92.17%, while the safety model accounts for merely 27% of the main model's parameters.

cs.RO↗

arXiv2024

An upper bound on the per-tile entropy of ribbon tilings

This paper considers $n$-ribbon tilings of general regions and their per-tile entropy (the binary logarithm of the number of tilings divided by the number of tiles). We show that the per-tile entropy is bounded above by $\log_2 n$. This bound improves the best previously known bounds of $n-1$ for general regions, and the asymptotic upper bound of $\log_2 (en)$ for growing rectangles, due to Chen and Kargin.

math.CO↗

arXiv2023

The Number of Ribbon Tilings for Strips

First, we consider order-$n$ ribbon tilings of an $M$-by-$N$ rectangle $R_{M,N}$ where $M$ and $N$ are much larger than $n$. We prove the existence of the growth rate $γ_n$ of the number of tilings and show that $γ_n \leq (n-1) \ln 2$. Then, we study a rectangle $R_{M,N}$ with fixed width $M=n$, called a strip. We derive lower and upper bounds on the growth rate $μ_n$ for strips as $ \ln n - 1 + o(1) \leq μ_n \leq \ln n $. Besides, we construct a recursive system which enables us to enumerate the order-$n$ ribbon tilings of a strip for all $n \leq 8$ and calculate the corresponding generating functions.

math.CO↗

arXiv2023

On enumeration and entropy of ribbon tilings

The paper considers ribbon tilings of large regions and their per-tile entropy (the logarithm of the number of tilings divided by the number of tiles). For tilings of general regions by ribbon tiles of length $n$, we give an upper bound on the per-tile entropy as $n - 1$. For growing rectangular regions, we prove the existence of the asymptotic per-tile entropy and show that it is bounded from below by $\log_2 (n/e)$ and from above by $\log_2(en)$. For growing generalized "Aztec Diamond'' regions and for growing "stair'' regions, the asymptotic per-tile entropy is calculated exactly as $1/2$ and $\log_2(n + 1) - 1$, respectively.

math.PR↗