arXiv subjects

Jianguo Huang

Publications and source records attributed to Jianguo Huang.

At least 37 records · Page 2Linked to original sources

arXiv2026

Conditional Performance Guarantee for Large Reasoning Models

Large reasoning models have shown strong performance through extended chain-of-thought reasoning, yet their computational cost remains significant. Probably approximately correct (PAC) reasoning provides statistical guarantees for efficient reasoning by adaptively switching between thinking and non-thinking models, but the guarantee holds only in the marginal case and does not provide exact conditional coverage. We propose G-PAC reasoning, a practical framework that provides PAC-style guarantees at the group level by partitioning the input space. We develop two instantiations: Group PAC (G-PAC) reasoning for known group structures and Clustered PAC (C-PAC) reasoning for unknown groupings. We prove that both G-PAC and C-PAC achieve group-conditional risk control, and that grouping can strictly improve efficiency over marginal PAC reasoning in heterogeneous settings. Our experiments on diverse reasoning benchmarks demonstrate that G-PAC and C-PAC successfully achieve group-conditional risk control while maintaining substantial computational savings.

cs.AI↗

arXiv2026

HyPAC: Cost-Efficient LLMs-Human Hybrid Annotation with PAC Error Guarantees

Data annotation often involves multiple sources with different cost-quality trade-offs, such as fast large language models (LLMs), slow reasoning models, and human experts. In this work, we study the problem of routing inputs to the most cost-efficient annotation source while controlling the labeling error on test instances. We propose \textbf{HyPAC}, a method that adaptively labels inputs to the most cost-efficient annotation source while providing distribution-free guarantees on annotation error. HyPAC calibrates two decision thresholds using importance sampling and upper confidence bounds, partitioning inputs into three regions based on uncertainty and routing each to the appropriate annotation source. We prove that HyPAC achieves the minimum expected cost with a probably approximately correct (PAC) guarantee on the annotation error, free of data distribution and pre-trained models. Experiments on common benchmarks demonstrate the effectiveness of our method, reducing the annotation cost by 78.51\% while tightly controlling the annotation error.

cs.LG↗

arXiv2026

Compression Tells Intelligence: Visual Coding, Visual Token Technology, and the Unification

"Compression Tells Intelligence", is supported by research in artificial intelligence, particularly concerning (multimodal) large language models (LLMs/MLLMs), where compression efficiency often correlates with improved model performance and capabilities. For compression, classical visual coding based on traditional information theory has developed over decades, achieving great success with numerous international industrial standards widely applied in multimedia (e.g., image/video) systems. Except that, the recent emergingvisual token technology of generative multi-modal large models also shares a similar fundamental objective like visual coding: maximizing semantic information fidelity during the representation learning while minimizing computational cost. Therefore, this paper provides a comprehensive overview of two dominant technique families first -- Visual Coding and Vision Token Technology -- then we further unify them from the aspect of optimization, discussing the essence of compression efficiency and model performance trade-off behind. Next, based on the proposed unified formulation bridging visual coding andvisual token technology, we synthesize bidirectional insights of themselves and forecast the next-gen visual codec and token techniques. Last but not least, we experimentally show a large potential of the task-oriented token developments in the more practical tasks like multimodal LLMs (MLLMs), AI-generated content (AIGC), and embodied AI, as well as shedding light on the future possibility of standardizing a general token technology like the traditional codecs (e.g., H.264/265) with high efficiency for a wide range of intelligent tasks in a unified and effective manner.

cs.CV↗

arXiv2025

MATAI: A Generalist Machine Learning Framework for Property Prediction and Inverse Design of Advanced Alloys

The discovery of advanced metallic alloys is hindered by vast composition spaces, competing property objectives, and real-world constraints on manufacturability. Here we introduce MATAI, a generalist machine learning framework for property prediction and inverse design of as-cast alloys. MATAI integrates a curated alloy database, deep neural network-based property predictors, a constraint-aware optimization engine, and an iterative AI-experiment feedback loop. The framework estimates key mechanical propertie, sincluding density, yield strength, ultimate tensile strength, and elongation, directly from composition, using multi-task learning and physics-informed inductive biases. Alloy design is framed as a constrained optimization problem and solved using a bi-level approach that combines local search with symbolic constraint programming. We demonstrate MATAI's capabilities on the Ti-based alloy system, a canonical class of lightweight structural materials, where it rapidly identifies candidates that simultaneously achieve lower density (<4.45 g/cm3), higher strength (>1000 MPa) and appreciable ductility (>5%) through only seven iterations. Experimental validation confirms that MATAI-designed alloys outperform commercial references such as TC4, highlighting the framework's potential to accelerate the discovery of lightweight, high-performance materials under real-world design constraints.

cond-mat.mtrl-sci↗

arXiv2025

Revisiting MLLM Token Technology through the Lens of Classical Visual Coding

Classical visual coding and Multimodal Large Language Model (MLLM) token technology share the core objective - maximizing information fidelity while minimizing computational cost. Therefore, this paper reexamines MLLM token technology, including tokenization, token compression, and token reasoning, through the established principles of long-developed visual coding area. From this perspective, we (1) establish a unified formulation bridging token technology and visual coding, enabling a systematic, module-by-module comparative analysis; (2) synthesize bidirectional insights, exploring how visual coding principles can enhance MLLM token techniques' efficiency and robustness, and conversely, how token technology paradigms can inform the design of next-generation semantic visual codecs; (3) prospect for promising future research directions and critical unsolved challenges. In summary, this study presents the first comprehensive and structured technology comparison of MLLM token and visual coding, paving the way for more efficient multimodal models and more powerful visual codecs simultaneously.

cs.CV↗

arXiv2025

Conformal Prediction Meets Long-tail Classification

Conformal Prediction (CP) is a popular method for uncertainty quantification that converts a pretrained model's point prediction into a prediction set, with the set size reflecting the model's confidence. Although existing CP methods are guaranteed to achieve marginal coverage, they often exhibit imbalanced coverage across classes under long-tail label distributions, tending to over cover the head classes at the expense of under covering the remaining tail classes. This under coverage is particularly concerning, as it undermines the reliability of the prediction sets for minority classes, even with coverage ensured on average. In this paper, we propose the Tail-Aware Conformal Prediction (TACP) method to mitigate the under coverage of the tail classes by utilizing the long-tail structure and narrowing the head-tail coverage gap. Theoretical analysis shows that it consistently achieves a smaller head-tail coverage gap than standard methods. To further improve coverage balance across all classes, we introduce an extension of TACP: soft TACP (sTACP) via a reweighting mechanism. The proposed framework can be combined with various non-conformity scores, and experiments on multiple long-tail benchmark datasets demonstrate the effectiveness of our methods.

cs.LG↗

arXiv2025

The 1st International Workshop on Disentangled Representation Learning for Controllable Generation (DRL4Real): Methods and Results

This paper reviews the 1st International Workshop on Disentangled Representation Learning for Controllable Generation (DRL4Real), held in conjunction with ICCV 2025. The workshop aimed to bridge the gap between the theoretical promise of Disentangled Representation Learning (DRL) and its application in realistic scenarios, moving beyond synthetic benchmarks. DRL4Real focused on evaluating DRL methods in practical applications such as controllable generation, exploring advancements in model robustness, interpretability, and generalization. The workshop accepted 9 papers covering a broad range of topics, including the integration of novel inductive biases (e.g., language), the application of diffusion models to DRL, 3D-aware disentanglement, and the expansion of DRL into specialized domains like autonomous driving and EEG analysis. This summary details the workshop's objectives, the themes of the accepted papers, and provides an overview of the methodologies proposed by the authors.

cs.LG↗

arXiv2025

Let's Revise Step-by-Step: A Unified Local Search Framework for Code Generation with LLMs

Large Language Models (LLMs) with inference-time scaling techniques show promise for code generation, yet face notable efficiency and scalability challenges. Construction-based tree-search methods suffer from rapid growth in tree size, high token consumption, and lack of anytime property. In contrast, improvement-based methods offer better performance but often struggle with uninformative reward signals and inefficient search strategies. In this work, we propose \textbf{ReLoc}, a unified local search framework which effectively performs step-by-step code revision. Specifically, ReLoc explores a series of local revisions through four key algorithmic components: initial code drafting, neighborhood code generation, candidate evaluation, and incumbent code updating, each of which can be instantiated with specific decision rules to realize different local search algorithms such as Hill Climbing (HC) or Genetic Algorithm (GA). Furthermore, we develop a specialized revision reward model that evaluates code quality based on revision distance to produce fine-grained preferences that guide the local search toward more promising candidates. Finally, our extensive experimental results demonstrate that our approach achieves superior performance across diverse code generation tasks, significantly outperforming both construction-based tree search as well as the state-of-the-art improvement-based code generation methods.

cs.CL↗

arXiv2025

An efficient numerical method for surface acoustic wave equations over unbounded domains

Surface acoustic wave (SAW) devices are widely used in modern communication equipment and SAW equations describe the critical physical processes of acoustic-electric conversion in SAW devices. It is very challenging to numerically solve such equations, since they are typically three dimensional problems defined on unbounded domains. In this paper, we first use the perfectly matched layer method to truncate the unbounded domain and then propose a finite element tearing and interconnecting algorithm for the truncated equations based on the periodic structure of the truncated domain. We also design an effective solver for the ill-conditioned linear system of the Lagrange multipliers arising from discretization. Several numerical results are performed to demonstrate the efficiency of the proposed algorithm.

math.NA↗

arXiv2025

C-Adapter: Adapting Deep Classifiers for Efficient Conformal Prediction Sets

Conformal prediction, as an emerging uncertainty quantification technique, typically functions as post-hoc processing for the outputs of trained classifiers. To optimize the classifier for maximum predictive efficiency, Conformal Training rectifies the training objective with a regularization that minimizes the average prediction set size at a specific error rate. However, the regularization term inevitably deteriorates the classification accuracy and leads to suboptimal efficiency of conformal predictors. To address this issue, we introduce \textbf{Conformal Adapter} (C-Adapter), an adapter-based tuning method to enhance the efficiency of conformal predictors without sacrificing accuracy. In particular, we implement the adapter as a class of intra order-preserving functions and tune it with our proposed loss that maximizes the discriminability of non-conformity scores between correctly and randomly matched data-label pairs. Using C-Adapter, the model tends to produce extremely high non-conformity scores for incorrect labels, thereby enhancing the efficiency of prediction sets across different coverage rates. Extensive experiments demonstrate that C-Adapter can effectively adapt various classifiers for efficient prediction sets, as well as enhance the conformal training method.

cs.LG↗

arXiv2025

Numerical analysis of variational-hemivariational inequalities with applications in contact mechanics

Variational-hemivariational inequalities are an important mathematical framework for nonsmooth problems. The framework can be used to study application problems from physical sciences and engineering that involve non-smooth and even set-valued relations, monotone or non-monotone, among physical quantities. Since no analytic solution formulas are expected for variational-hemivariational inequalities from applications, numerical methods are needed to solve the problems. This paper focuses on numerical analysis of variational-hemivariational inequalities, reporting new results as well as surveying some recent published results in the area. A general convergence result is presented for Galerkin solutions of the inequalities under minimal solution regularity conditions available from the well-posedness theory, and Céa's inequalities are derived for error estimation of numerical solutions. The finite element method and the virtual element method are taken as examples of numerical methods, optimal order error estimates for the linear element solutions are derived when the methods are applied to solve three representative contact problems under certain solution regularity assumptions. Numerical results are presented to show the performance of both the finite element method and the virtual element method, including numerical convergence orders of the numerical solutions that match the theoretical predictions.

math.NA↗

arXiv2025

Similarity-Navigated Conformal Prediction for Graph Neural Networks

Graph Neural Networks have achieved remarkable accuracy in semi-supervised node classification tasks. However, these results lack reliable uncertainty estimates. Conformal prediction methods provide a theoretical guarantee for node classification tasks, ensuring that the conformal prediction set contains the ground-truth label with a desired probability (e.g., 95%). In this paper, we empirically show that for each node, aggregating the non-conformity scores of nodes with the same label can improve the efficiency of conformal prediction sets while maintaining valid marginal coverage. This observation motivates us to propose a novel algorithm named Similarity-Navigated Adaptive Prediction Sets (SNAPS), which aggregates the non-conformity scores based on feature similarity and structural neighborhood. The key idea behind SNAPS is that nodes with high feature similarity or direct connections tend to have the same label. By incorporating adaptive similar nodes information, SNAPS can generate compact prediction sets and increase the singleton hit ratio (correct prediction sets of size one). Moreover, we theoretically provide a finite-sample coverage guarantee of SNAPS. Extensive experiments demonstrate the superiority of SNAPS, improving the efficiency of prediction sets and singleton hit ratio while maintaining valid coverage.

cs.LG↗

arXiv2024

A robust $C^0$ interior penalty method for a gradient-elastic Kirchhoff plate model

This paper is devoted to proposing and analyzing a robust $C^0$ interior penalty method for a gradient-elastic Kirchhoff plate (GEKP) model over a convex polygon. The numerical method is obtained by combining the triangular Hermite element and a $C^0$ interior penalty method, which can avoid the use of higher order shape functions or macroelements. Next, a robust regularity estimate is established for the GEKP model based on our earlier result for a triharmonic equation on a convex polygon. Furthermore, some local lower bound estimates of the a posteriori error analysis are established. These together with an enriching operator and its error estimates lead to a Céa-like lemma. Thereby, the optimal error estimates are achieved, which are also robust with respect to the small size parameter. In addition, it is proved that this numerical method is convergent without any additional regularity assumption for the exact solution. Some numerical experiments are performed to verify the theoretical findings.

math.NA↗

arXiv2024

Does confidence calibration improve conformal prediction?

Conformal prediction is an emerging technique for uncertainty quantification that constructs prediction sets guaranteed to contain the true label with a predefined probability. Previous works often employ temperature scaling to calibrate classifiers, assuming that confidence calibration benefits conformal prediction. However, the specific impact of confidence calibration on conformal prediction remains underexplored. In this work, we make two key discoveries about the impact of confidence calibration methods on adaptive conformal prediction. Firstly, we empirically show that current confidence calibration methods (e.g., temperature scaling) typically lead to larger prediction sets in adaptive conformal prediction. Secondly, by investigating the role of temperature value, we observe that high-confidence predictions can enhance the efficiency of adaptive conformal prediction. Theoretically, we prove that predictions with higher confidence result in smaller prediction sets on expectation. This finding implies that the rescaling parameters in these calibration methods, when optimized with cross-entropy loss, might counteract the goal of generating efficient prediction sets. To address this issue, we propose Conformal Temperature Scaling (ConfTS), a variant of temperature scaling with a novel loss function designed to enhance the efficiency of prediction sets. This approach can be extended to optimize the parameters of other post-hoc methods of confidence calibration. Extensive experiments demonstrate that our method improves existing adaptive conformal prediction methods in classification tasks, especially with LLMs.

cs.LG↗

arXiv2024

A $C^0$-continuous nonconforming virtual element method for linear strain gradient elasticity

A robust $C^0$-continuous nonconforming virtual element method (VEM) is developed for a boundary value problem arising from strain gradient elasticity in two dimensions, with the family of polygonal meshes satisfying a very general geometric assumption given in Brezzi et al. (2009) and Chen and Huang (2018). The stability condition of the VEMs is derived by establishing Korn-type inequalities and inverse inequalities. Some crucial commutative relations for locking-free analysis as in elastic problems are derived. The sharp and uniform error estimates with respect to both the microscopic parameter and the Lamé coefficient are achieved in the lowest-order case, which is also verified by numerical results.

math.NA↗

arXiv2024

A Parareal exponential integrator finite element method for linear parabolic equations

In this paper, for solving a class of linear parabolic equations in rectangular domains, we have proposed an efficient Parareal exponential integrator finite element method. The proposed method first uses the finite element approximation with continuous multilinear rectangular basis function for spatial discretization, and then takes the Runge-Kutta approach accompanied with Parareal framework for time integration of the resulting semi-discrete system to produce parallel-in-time numerical solution. Under certain regularity assumptions, fully-discrete error estimates in $L^2$-norm are derived for the proposed schemes with random interpolation nodes. Moreover, a fast solver can be provided based on tensor product spectral decomposition and fast Fourier transform (FFT), since the mass and coefficient matrices of the proposed method can be simultaneously diagonalized with an orthogonal matrix. A series of numerical experiments in various dimensions are also presented to validate the theoretical results and demonstrate the excellent performance of the proposed method.

math.NA↗

arXiv2024

Exponential integrator Fourier Galerkin methods for semilinear parabolic equations

In this paper, in order to improve the spatial accuracy, the exponential integrator Fourier Galerkin method (EIFG) is proposed for solving semilinear parabolic equations in rectangular domains. In this proposed method, the spatial discretization is first carried out by the Fourier-based Galerkin approximation, and then the time integration of the resulting semi-discrete system is approximated by the explicit exponential Runge-Kutta approach, which leads to the fully-discrete numerical solution. With certain regularity assumptions on the model problem, error estimate measured in $H^2$-norm is explicitly derived for EIFG method with two RK stages. Several two and three dimensional examples are shown to demonstrate the excellent performance of EIFG method, which are coincident to the theoretical results.

math.NA↗

arXiv2024

Adaptive neural network basis methods for partial differential equations with low-regular solutions

This paper aims to devise an adaptive neural network basis method for numerically solving a second-order semilinear partial differential equation (PDE) with low-regular solutions in two/three dimensions. The method is obtained by combining basis functions from a class of shallow neural networks and the resulting multi-scale analogues, a residual strategy in adaptive methods and the non-overlapping domain decomposition method. At the beginning, in view of the solution residual, we partition the total domain $Ω$ into $K+1$ non-overlapping subdomains, denoted respectively as $\{Ω_k\}_{k=0}^K$, where the exact solution is smooth on subdomain $Ω_{0}$ and low-regular on subdomain $Ω_{k}$ ($1\le k\le K$). Secondly, the low-regular solutions on different subdomains $Ω_{k}$~($1\le k\le K$) are approximated by neural networks with different scales, while the smooth solution on subdomain $Ω_0$ is approximated by the initialized neural network. Thirdly, we determine the undetermined coefficients by solving the linear least squares problems directly or the nonlinear least squares problem via the Gauss-Newton method. The proposed method can be extended to multi-level case naturally. Finally, we use this adaptive method for several peak problems in two/three dimensions to show its high-efficient computational performance.

math.NA↗