arXiv subjects

Xinlei Wang

Publications and source records attributed to Xinlei Wang.

At least 19 recordsLinked to original sources

arXiv2026

SnapBench: Benchmarking Snap-and-Ask Multimodal Retrieval for Mobile Interactions

Mobile AI acts as a visual oracle, empowering users to snap a picture of something and ask for information. Snap-and-ask retrieval is now one of the most common entry points for mobile AI, yet photos are often blurry, while text questions may be short or mistyped. Existing benchmarks only test on clean inputs or do not isolate paired robustness in snap-and-ask retrieval. Therefore, we introduce SnapBench, the first paired benchmark for robust snap-and-ask multimodal retrieval, spanning 1,145 queries, 9,085 gallery items under 53 controlled corruption conditions with human annotations. We evaluate 16 multimodal retrievers, covering dual-tower encoders and embedding-based VLMs. Results show that image corruptions substantially degrade retrieval, while text corruptions mainly affect text-only retrieval and have limited impact on joint retrieval. Clean image-only retrieval often outperforms joint retrieval, indicating the coarse-text drag and the lack of cross-modal fallback under noisy inputs. SnapBench provides a controlled testbed for evaluating robust retrieval in snap-and-ask scenarios. We further propose MOOR (Modality-anchored, Outlier-aware, Optimal Reweighting), a simple adaptive fusion approach, highlighting the need for reliability-aware modality calibration in snap-and-ask retrieval.

cs.CV↗

arXiv2026

RePro: Proof-Verified Benchmark Rewriting for Reliable Evaluation of LLM Mathematical Problem Solving

Data contamination undermines the reliable evaluation of large language models (LLMs) on mathematical problem solving. While rewriting-based evaluation mitigates memorization, existing methods lack guarantees of problem validity and answer correctness. We propose Proof-Verified Benchmark Rewriting (RePro), the first framework to integrate Lean-oriented neural automated theorem provers (ATPs) into benchmark rewriting, which rewrites problems and regenerates answers with correctness ensured by Lean-verified proofs. Experiments on GSM8K and MATH show that RePro's retained rewritten instances achieve 100% well-definedness, feasibility, and answer correctness, while existing methods still produce invalid or incorrect instances. Moreover, several models exhibit accuracy drops on proof-verified rewritten benchmarks, suggesting that their performance is sensitive to surface-level and structural variations and may partly reflect memorization effects. Our source code and data are available at https://github.com/AI4Engi/RePro.

cs.CL↗

arXiv2026

Emission-Forecasting-Based Spatial-Temporal Carbon Response: A Multi-Agent Attention-Enhanced Deep Learning Framework

As a major contributor to carbon emissions, the decarbonization of power systems has garnered significant societal attention. Nodal carbon intensity (NCI), a critical factor in carbon-oriented demand response, has traditionally been determined through ex-post calculations. However, this ex-post approach introduces latency in low-carbon dispatch. To address this, this paper presents a proactive ex-ante spatial-temporal carbon response framework. At its core, we develop a novel deep learning-based hierarchical design, enhanced by a dual-stage attention mechanism and a large language model (LLM)-based multi-agent cooperation system, to accurately forecast day-ahead NCI. This design effectively mitigates the impact of renewable energy uncertainty and enhances predictive resilience. On the demand side, the framework proposes a spatial-temporal carbon scheduling model that integrates geographically dispatchable loads (GDLs), including mobile energy storage systems (MESSs) and distributed data centers (DDCs). Leveraging high-accuracy day-ahead NCI predictions, the framework can effectively reduce system emissions by quickly responding to carbon intensity fluctuations. The proposed framework is tested on the modified IEEE 33-bus system. According to the simulation results, the impacts of proposed framework on dispatching latency and emission outcomes are analyzed. The results demonstrate that under a one-hour reduction in carbon scheduling latency, the proposed model and methodology can achieve over 30% emission reduction. This research breaks through the limitations of passive carbon accounting, advancing toward proactive carbon management. It offers an intelligent solution that accelerates the transition to cleaner power systems while directly supporting sustainable production goals.

eess.SY↗

arXiv2026

CoLaDAG: Compositional Latent Log-ratio DAG Analysis of the Gut Microbiome under Silver Nanoparticle Exposure

Directed network analysis of microbiome counts is complicated by compositional sampling, high dimensionality, and limited biological replication. We present CoLaDAG, a fixed-reference latent additive log-ratio (ALR) estimator for generating sparse directed conditional-dependence hypotheses from compositional counts. The method combines a multinomial observation model, a working linear Gaussian structural equation model, nonconvex DC-ADMM optimization, and post-estimation thresholding with greedy acyclic projection. Under simulations aligned with this observation model, CoLaDAG obtained the largest mean exact-direction Matthews correlation and the smallest mean false discovery rate among the evaluated implementations; performance deteriorated under continuous-data and dropout misspecification. In the 12-mouse silver-nanoparticle (AgNP) case study, the 58-node fitted graph was sensitive to block resampling and ALR reference choice: 60 of 284 primary edges attained a mouse-block selection frequency of at least 0.60. The reported orientations and dose-stratified slopes are exploratory, coordinate-specific hypotheses rather than identified causal or exposure effects. The leading stable relations prioritize anaerobic gut taxa for targeted abundance, metabolite, and perturbation studies, but do not establish cross-feeding or toxicological mechanisms.

stat.AP↗

arXiv2026

Bayesian Donor Set Selection in Synthetic Controls

The Synthetic Control Method (SCM) is a widely used approach for assessing the effects of interventions by constructing a synthetic counterfactual using a donor set of untreated units. However, the effectiveness of SCM heavily relies on the careful selection of an appropriate donor set. In this paper, we propose a Bayesian hierarchical model that performs donor set selection while preserving the standard SCM simplex constraint on donor weights. Unlike approaches that assume a fixed donor set, our model allows for the simultaneous estimation of the synthetic control weights and the active donor set. By using a hierarchical Gamma-Bernoulli construction for the donor weights, the proposed model assigns posterior mass to simplex faces and allows exact zero weights for excluded donors. We establish a posterior donor-set consistency result under a simplified pre-intervention model. Through numerical simulations, we show that our model improves donor recovery and weight estimation when the donor pool contains irrelevant or weakly related units, while remaining competitive in full-donor settings. Finally, we apply our model to the GDP trajectory of West Germany, illustrating its practical applicability. Our findings suggest that incorporating donor set selection offers a more parsimonious and flexible extension of existing Bayesian synthetic control methods.

stat.ME↗

arXiv2026

Rendering-Aware Bayesian 3D Gaussian Splatting with Native Uncertainty and Adaptive Complexity Control

3D Gaussian splatting (3DGS) is a strong representation for real-time novel-view synthesis, but its standard training pipeline relies on point estimates and hand-tuned heuristics, providing no native uncertainty or principled complexity control. This is most limiting under sparse views or fixed acquisition budgets, where a model must identify weakly supported geometry and select informative views. We introduce a rendering-aware Bayesian 3DGS framework that tracks Gaussian geometry with a Normal-Inverse-Wishart posterior over means and covariances using renderer-derived surrogate summaries. An optional Dirichlet-process extension adds a probabilistic component-usage signal, and the training schedule makes the closed-form versus approximate inference boundary explicit. Re-rendering posterior geometry samples yields native predictive uncertainty for interval calibration and active view selection. In a fixed-budget 16-to-32 active-view task, native NIW acquisition improves PSNR by +0.453 dB and LPIPS by -0.0146 over a scoring-only 3-member standard-ensemble baseline, winning 29/39 scene-seed pairs and 10/13 scene means; it also improves over PPU-style (+0.355 dB) and NIW-proxy (+0.401 dB) acquisition. NIW native intervals reduce 95% coverage error by about 17x relative to a shared proxy (0.046 vs. 0.796) and are about 10x closer to nominal coverage than a 3-member deep ensemble (0.047 vs. 0.454) at roughly one-third the training cost. As a reconstruction compatibility check, paired NIW-vs-standard analysis over 39 scene-seed runs yields +0.030 dB PSNR with 1.6% additional training time. These results position Bayesian 3DGS as a practical probabilistic scene representation for decision-facing tasks such as active view selection.

cs.CV↗

arXiv2026

Hugging Carbon: Quantifying the Training Carbon Emissions of AI Models at Scale

The scaling-law era has transformed artificial intelligence (AI) from research into a global industry, but its rapid growth also raises concerns over energy usage, carbon emissions, and environmental sustainability. Unlike traditional sectors, the AI industry still lacks systematic carbon accounting methods that support large-scale estimates without reproducing the original training process. This leaves open questions about how large the problem is today and how large it might be in the near future. Given its central role in hosting open-source AI models, the Hugging Face (HF) platform provides a large-scale and publicly accessible corpus for carbon accounting. We estimate aggregate training emissions of HF open-source models using available emissions, energy, compute, and model metadata. To address uneven disclosure quality, we introduce a tiered approach to handle incomplete metadata, supported by empirical regressions that assess estimation reliability. We further introduce AI training carbon intensity (ATCI, emissions per compute), a metric to assess the sustainability efficiency of model training. Our results show that training the most popular open-source models (with over 5,000 downloads) has already resulted in approximately $6.0\times10^4$ metric tons of carbon emissions. Overall, this paper provides a scalable, empirically grounded framework for estimating training emissions from incomplete disclosures and informing future carbon reporting standards in the AI industry. Data and code are available at https://github.com/insait-institute/HuggingCarbon.

cs.CY↗

arXiv2026

Uncertainty-Aware Ideal Point Estimation via Variational EM

Roll-call data analysis aims to estimate legislators' ideal points and quantify the associated uncertainty. Existing approaches either rely on Bayesian methods implemented via Markov chain Monte Carlo sampling or focus primarily on point estimation, with uncertainty typically assessed through resampling procedures such as the bootstrap. Consequently, the computational burden of these approaches can become substantial when applied to large roll-call datasets. To address this challenge, we propose a computationally efficient likelihood method for estimating ideal points and their standard errors. Leveraging the Pólya--Gamma identity, we develop a variational expectation--maximization algorithm for estimating ideal points and introduce a variational Louis' method to approximate the observed Fisher information for standard error estimation. Numerical studies and applications to U.S. congressional roll-call data demonstrate that the proposed method produces accurate ideal point estimates and reliable standard errors while being substantially more computationally efficient than existing approaches.

stat.ME↗

arXiv2026

EngiAgent: Fully Connected Coordination of LLM Agents for Solving Open-ended Engineering Problems with Feasible Solutions

Engineering problem solving is central to real-world decision-making, requiring mathematical formulations that not only represent complex problems but also produce feasible solutions under data and physical constraints. Unlike mathematical problem solving, which operates on predefined formulations, engineering tasks demand open-ended analysis, feasibility-driven modeling, and iterative refinement. Although large language models (LLMs) have shown strong capabilities in reasoning and code generation, they often fail to ensure feasibility, which limits their applicability to engineering problem solving. To address this challenge, we propose EngiAgent, a multi-agent system with a fully connected coordinator that simulates expert workflows through specialized agents for problem analysis, modeling, verification, solving, and solution evaluation. The fully connected coordinator enables flexible feedback routing, overcoming the rigidity of prior pipeline-based reflection methods and ensuring feasibility at every stage of the process. This design not only improves robustness to diverse failure cases such as data extraction errors, constraint inconsistencies, and solver failures, but also enhances the overall quality of problem solving. Empirical results across four representative domains demonstrate that EngiAgent achieves substantial improvements in feasibility compared to prior approaches, establishing a new paradigm for feasibility-oriented engineering problem solving with LLMs. Our source code and data are available at https://github.com/AI4Engi/EngiAgent.

cs.AI↗

arXiv2026

EngiBench: A Benchmark for Evaluating Large Language Models on Engineering Problem Solving

Large language models (LLMs) have shown strong performance on mathematical reasoning under well-defined conditions. However, real-world engineering problems involve uncertainty, context, and open-ended settings that extend beyond symbolic computation. Existing benchmarks largely focus on well-defined or abstract reasoning and therefore fail to capture these complexities. We introduce EngiBench, a hierarchical benchmark designed to evaluate LLMs on solving engineering problems. It spans three levels of increasing difficulty (foundational knowledge retrieval, contextual reasoning, and open-ended modeling) and covers diverse engineering subfields. To facilitate a deeper understanding of model performance, we systematically rewrite each problem into three controlled variants (perturbed, knowledge-enhanced, and math abstraction), enabling us to separately evaluate the model's robustness, domain-specific knowledge, and mathematical reasoning abilities. Experimental results show clear performance stratification across difficulty levels: model accuracy declines with task complexity, degrades under minor perturbations, and remains substantially below human performance on high-level engineering tasks. These findings reveal that current LLMs still lack the high-level reasoning needed for real-world engineering, highlighting the need for future models with deeper and more reliable problem-solving capabilities. Our source code and data are available at https://github.com/AI4Engi/EngiBench.

cs.AI↗

arXiv2026

CropCraft: Complete Structural Characterization of Crop Plants From Images

The ability to automatically build 3D digital twins of plants from images has countless applications in agriculture, environmental science, robotics, and other fields. However, current 3D reconstruction methods fail to recover complete shapes of plants due to heavy occlusion and complex geometries. In this work, we present a novel method for 3D modeling of agricultural crops based on optimizing a parametric model of plant morphology via inverse procedural modeling. Our method first estimates depth maps by fitting a neural radiance field and then optimizes a specialized loss to estimate morphological parameters that result in consistent depth renderings. The resulting 3D model is complete and biologically plausible. We validate our method on a dataset of real images of agricultural fields, and demonstrate that the reconstructed canopies can be used for a variety of monitoring and simulation applications.

cs.CV↗

arXiv2025

Empirical Bayes Method for Large Scale Multiple Testing with Heteroscedastic Errors

In this paper, we address the normal mean inference problem, which involves testing multiple means of normal random variables with heteroscedastic variances. Most existing empirical Bayes methods for this setting are developed under restrictive assumptions, such as the scaled inverse-chi-squared prior for variances and unimodality for the non-null mean distribution. However, when either of these assumptions is violated, these methods often fail to control the false discovery rate (FDR) at the target level or suffer from a substantial loss of power. To overcome these limitations, we propose a new empirical Bayes method, gg-Mix, which assumes only independence between the normal means and variances, without imposing any structural restrictions on their distributions. We thoroughly evaluate the FDR control and power of gg-Mix through extensive numerical studies and demonstrate its superior performance compared to existing methods. Finally, we apply gg-Mix to three real data examples to further illustrate the practical advantages of our approach.

stat.ME↗

arXiv2025

Transitional Dynamics: Unveiling the Coexistence and Interplay of Type-B and Type-C QPOs in MAXI J1348-630

Based on broadband timing analysis of Insight-HXMT and NICER data from the 2019 outburst of the black hole X-ray binary (BHXRB) MAXI J1348-630, we report the detection of the coexistence and competitive interplay between type-C and type-B quasi-periodic oscillations (QPOs). Specifically, the two QPO types were detected simultaneously but exhibited distinct energy dependencies: the type-C QPO was dominant in the hard X-ray band (10-30 keV), while the type-B QPO was more prominent in the soft X-ray band (1-10 keV). Further analysis reveals that the emergence of the type-C QPO suppresses the strength of the type-B QPO, particularly above 10 keV. Tracking the detailed evolution of these QPOs, we find that the weakening of the type-B QPO coincides with the strengthening of the type-C QPO, indicating a competitive interaction between them. These findings reveal a more complex relationship between type-B and type-C QPOs than previously recognized, suggesting they originate from distinct but interacting physical components within the accretion flow and/or jet, and providing new insights into the dynamics of accretion state transitions.

astro-ph.HE↗

arXiv2025

Demeter: A Parametric Model of Crop Plant Morphology from the Real World

Learning 3D parametric shape models of objects has gained popularity in vision and graphics and has showed broad utility in 3D reconstruction, generation, understanding, and simulation. While powerful models exist for humans and animals, equally expressive approaches for modeling plants are lacking. In this work, we present Demeter, a data-driven parametric model that encodes key factors of a plant morphology, including topology, shape, articulation, and deformation into a compact learned representation. Unlike previous parametric models, Demeter handles varying shape topology across various species and models three sources of shape variation: articulation, subcomponent shape variation, and non-rigid deformation. To advance crop plant modeling, we collected a large-scale, ground-truthed dataset from a soybean farm as a testbed. Experiments show that Demeter effectively synthesizes shapes, reconstructs structures, and simulates biophysical processes. Code and data is available at https://tianhang-cheng.github.io/Demeter/.

cs.CV↗

arXiv2025

Toward Reasoning-Centric Time-Series Analysis

Traditional time series analysis has long relied on pattern recognition, trained on static and well-established benchmarks. However, in real-world settings -- where policies shift, human behavior adapts, and unexpected events unfold -- effective analysis must go beyond surface-level trends to uncover the actual forces driving them. The recent rise of Large Language Models (LLMs) presents new opportunities for rethinking time series analysis by integrating multimodal inputs. However, as the use of LLMs becomes popular, we must remain cautious, asking why we use LLMs and how to exploit them effectively. Most existing LLM-based methods still employ their numerical regression ability and ignore their deeper reasoning potential. This paper argues for rethinking time series with LLMs as a reasoning task that prioritizes causal structure and explainability. This shift brings time series analysis closer to human-aligned understanding, enabling transparent and context-aware insights in complex real-world environments.

cs.AI↗

arXiv2025

Carbon Disclosure Effect, Corporate Fundamentals, and Net-zero Emission Target: Evidence from China

In response to China's national carbon neutrality goals, this study examines how corporate carbon emissions disclosure affects the financial performance of Chinese A-share listed companies. Leveraging artificial intelligence tools, including natural language processing, we analyzed emissions disclosures for 4,336 companies from 2017 to 2022. The research demonstrates that high-quality carbon disclosure positively impacts financial performance with higher stock returns, improved return on equity, increased Tobin's Q ratio, and reduced stock price volatility. Our findings underscore the emerging importance of carbon transparency in financial markets, highlighting how environmental reporting can serve as a strategic mechanism to create corporate value and adapt to climate change.

econ.GN↗

arXiv2025

Universal Magnetocaloric Effect near Quantum Critical Point of Magnon Bose-Einstein Condensation

Bose-Einstein condensation (BEC), a macroscopic quantum phenomenon arising from phase coherence and bosonic statistics, has been realized in quantum magnets. Here, we report the observation of a universal magnetocaloric effect (MCE) near a BEC quantum critical point (QCP) in copper sulfate crystal ($CuSO_4 \cdot 5H_2O$). By conducting magnetocaloric and nuclear magnetic resonance measurements, we uncover a field-driven BEC QCP, evidenced by the universal scaling law $T_c \propto (B_c - B)^{2/3}$ and the perfect data collapse of the magnetic Grüneisen ratio. Thermal excitation triggers a dimensional crossover to a 1D quantum-critical regime, where the MCE scaling strictly matches the universality class of 1D Fermi gases. Notably, the quantum-critical MCE enables cooling down to 12.8 mK without helium-3, with very fast thermal relaxation rate that is critical for high cooling power. This work demonstrates the universal MCE in magnon BEC systems, using a common copper sulfate compound as a paradigmatic example, and paves the way for next-generation sub-Kelvin cooling.

cond-mat.str-el↗

arXiv2025

Inference for Error-Prone Count Data: Estimation under a Binomial Convolution Framework

Measurement error in count data is common but underexplored in the literature, particularly in contexts where observed scores are bounded and arise from discrete scoring processes. Motivated by applications in oral reading fluency assessment, we propose a binomial convolution framework that extends binary misclassification models to settings where only the aggregate number of correct responses is observed, and errors may involve both overcounting and undercounting the number of events. The model accommodates distinct true positive and true negative accuracy rates and preserves the bounded nature of the data. Assuming the availability of both contaminated and error-free scores on a subset of items, we develop and compare three estimation strategies: maximum likelihood estimation (MLE), linear regression, and generalized method of moments (GMM). Extensive simulations show that MLE is most accurate when the model is correctly specified but is computationally intensive and less robust to misspecification. Regression is simple and stable but less precise, while GMM offers a compromise in model dependence, though it is sensitive to outliers. In practice, this framework supports improved inference in unsupervised settings where contaminated scores serve as inputs to downstream analyses. By quantifying accuracy rates, the model enables score corrections even when no specific outcome is yet defined. We demonstrate its utility using real oral reading fluency data, comparing human and AI-generated scores. Findings highlight the practical implications of estimator choice and underscore the importance of explicitly modeling asymmetric measurement error in count data.

stat.ME↗