arXiv subjects

Kai Wang

Publications and source records attributed to Kai Wang.

At least 253 records · Page 14Linked to original sources

arXiv2025

Neural Index Policies for Restless Multi-Action Bandits with Heterogeneous Budgets

Restless multi-armed bandits (RMABs) provide a scalable framework for sequential decision-making under uncertainty, but classical formulations assume binary actions and a single global budget. Real-world settings, such as healthcare, often involve multiple interventions with heterogeneous costs and constraints, where such assumptions break down. We introduce a Neural Index Policy (NIP) for multi-action RMABs with heterogeneous budget constraints. Our approach learns to assign budget-aware indices to arm--action pairs using a neural network, and converts them into feasible allocations via a differentiable knapsack layer formulated as an entropy-regularized optimal transport (OT) problem. The resulting model unifies index prediction and constrained optimization in a single end-to-end differentiable framework, enabling gradient-based training directly on decision quality. The network is optimized to align its induced occupancy measure with the theoretical upper bound from a linear programming relaxation, bridging asymptotic RMAB theory with practical learning. Empirically, NIP achieves near-optimal performance within 5% of the oracle occupancy-measure policy while strictly enforcing heterogeneous budgets and scaling to hundreds of arms. This work establishes a general, theoretically grounded, and scalable framework for learning index-based policies in complex resource-constrained environments.

cs.LG↗

arXiv2025

GenColorBench: A Color Evaluation Benchmark for Text-to-Image Generation Models

Recent years have seen impressive advances in text-to-image generation, with image generative or unified models producing high-quality images from text. Yet these models still struggle with fine-grained color controllability, often failing to accurately match colors specified in text prompts. While existing benchmarks evaluate compositional reasoning and prompt adherence, none systematically assess color precision. Color is fundamental to human visual perception and communication, critical for applications from art to design workflows requiring brand consistency. However, current benchmarks either neglect color or rely on coarse assessments, missing key capabilities such as interpreting RGB values or aligning with human expectations. To this end, we propose GenColorBench, the first comprehensive benchmark for text-to-image color generation, grounded in color systems like ISCC-NBS and CSS3/X11, including numerical colors which are absent elsewhere. With 44K color-focused prompts covering 400+ colors, it reveals models' true capabilities via perceptual and automated assessments. Evaluations of popular text-to-image models using GenColorBench show performance variations, highlighting which color conventions models understand best and identifying failure modes. Our GenColorBench assessments will guide improvements in precise color generation. The benchmark will be made public upon acceptance.

cs.CV↗

arXiv2025

Zero-Dimensional Stacking Domains Enable Strong-Ductile Synergy in Additive Manufactured Titanium

Alloying by addition of oxygen interstitials during additive manufacturing provides new routes to strengthen and toughen metals and alloys. The underlying mechanisms by which such interstitial atoms lead to enhanced properties remain, however, unclear, not least due a lack of quantitative atomic-scale models linking microstructure to properties. Here using quasi-3D imaging based on multi-slice electron ptychography, we reveal the importance of a new type of interstitial-character lattice defect, namely zero-dimensional stacking domains (ZDSDs), present in high density in AM-processed oxygen-modulated pure titanium. These ZDSDs promote slip diversity, and support intense work hardening, enabling a three-fold enhancement in both strength and ductility in Ti-0.45O compared to conventional pure Ti. The work demonstrates the potential for using interstitial solutes to enhance mechanical properties in a range of critical engineering alloys.

cond-mat.mtrl-sci↗

arXiv2025

From Cradle to Cane: A Two-Pass Framework for High-Fidelity Lifespan Face Aging

Face aging has become a crucial task in computer vision, with applications ranging from entertainment to healthcare. However, existing methods struggle with achieving a realistic and seamless transformation across the entire lifespan, especially when handling large age gaps or extreme head poses. The core challenge lies in balancing age accuracy and identity preservation--what we refer to as the Age-ID trade-off. Most prior methods either prioritize age transformation at the expense of identity consistency or vice versa. In this work, we address this issue by proposing a two-pass face aging framework, named Cradle2Cane, based on few-step text-to-image (T2I) diffusion models. The first pass focuses on solving age accuracy by introducing an adaptive noise injection (AdaNI) mechanism. This mechanism is guided by including prompt descriptions of age and gender for the given person as the textual condition. Also, by adjusting the noise level, we can control the strength of aging while allowing more flexibility in transforming the face. However, identity preservation is weakly ensured here to facilitate stronger age transformations. In the second pass, we enhance identity preservation while maintaining age-specific features by conditioning the model on two identity-aware embeddings (IDEmb): SVR-ArcFace and Rotate-CLIP. This pass allows for denoising the transformed image from the first pass, ensuring stronger identity preservation without compromising the aging accuracy. Both passes are jointly trained in an end-to-end way. Extensive experiments on the CelebA-HQ test dataset, evaluated through Face++ and Qwen-VL protocols, show that our Cradle2Cane outperforms existing face aging methods in age accuracy and identity consistency. Code is available at https://github.com/byliutao/Cradle2Cane.

cs.CV↗

arXiv2025

The Bigfoot: A footprint of a Coma cluster progenitor at z=3.98

Protoclusters, galaxy clusters' high redshift progenitors, hold the keys to understanding the formation and evolution of clusters and their member galaxies. However, their cosmological distances and spatial extensions (tens of Mpc) have inhibited complete mapping of their structure and constituent galaxies, which is key to robustly linking protoclusters to their descendants. Here we report the discovery of the Bigfoot, a tridimensional structure at $z = 3.98$ including 11 subgroups traced by 55 (700) spectroscopic (photometric) redshifts with JWST, extending over $15\times 37$ $\times 49{\rm{cMpc^3}}$ in the PRIMER-UDS field. Bigfoot's large-scale and mass function of member galaxies closely match constrained simulations' predictions for the progenitors of today's most massive clusters (${M_0} > 10^{15} {M_{_ \odot }}$). All subgroups with ${M_{\rm{h}}} > {10^{12.5}}{M_{_ \odot }}$ exhibit enhanced fractions of massive galaxies ($>{10^{10.0} {M_{_ \odot }}}$) compared to lower-mass halos and the field, demonstrating the accelerated formation of massive galaxies in massive halos. The presence of this massive protocluster with a large central halo (${10^{13.0} {M_{_ \odot }}}$) in a JWST deep field bears important cosmological implication that favors high ${σ_8}$ of PLANCK cosmology over low-redshift probes.

astro-ph.GA↗

arXiv2025

One-Step Flow Policy Mirror Descent

Diffusion policies have achieved great success in online reinforcement learning (RL) due to their strong expressive capacity. However, the inference of diffusion policy models relies on a slow iterative sampling process, which limits their responsiveness. To overcome this limitation, we propose Flow Policy Mirror Descent (FPMD), an online RL algorithm that enables 1-step sampling during flow policy inference. Our approach exploits a theoretical connection between the distribution variance and the discretization error of single-step sampling in straight interpolation flow matching models, and requires no extra distillation or consistency training. We present two algorithm variants based on rectified flow policy and MeanFlow policy, respectively. Extensive empirical evaluations on MuJoCo and visual DeepMind Control Suite benchmarks demonstrate that our algorithms show strong performance comparable to diffusion policy baselines while requiring orders of magnitude less computational cost during inference.

cs.LG↗

arXiv2025

Diffusion-DFL: Decision-focused Diffusion Models for Stochastic Optimization

Decision-focused learning (DFL) integrates predictive modeling and optimization by training predictors to optimize the downstream decision target rather than merely minimizing prediction error. To date, existing DFL methods typically rely on deterministic point predictions, which are often insufficient to capture the intrinsic stochasticity of real-world environments. To address this challenge, we propose the first diffusion-based DFL approach, which trains a diffusion model to represent the distribution of uncertain parameters and optimizes the decision by solving a stochastic optimization with samples drawn from the diffusion model. Our contributions are twofold. First, we formulate diffusion DFL using the reparameterization trick, enabling end-to-end training through diffusion. While effective, it is memory and compute-intensive due to the need to differentiate through the diffusion sampling process. Second, we propose a lightweight score function estimator that uses only several forward diffusion passes and avoids backpropagation through the sampling. This follows from our results that backpropagating through stochastic optimization can be approximated by a weighted score function formulation. We empirically show that our diffusion DFL approach consistently outperforms strong baselines in decision quality. The source code for all experiments is available at the project repository: https://github.com/GT-KOALA/Diffusion_DFL.

cs.LG↗

arXiv2025

Covariances for Free: Exploiting Mean Distributions for Training-free Federated Learning

Using pre-trained models has been found to reduce the effect of data heterogeneity and speed up federated learning algorithms. Recent works have explored training-free methods using first- and second-order statistics to aggregate local client data distributions at the server and achieve high performance without any training. In this work, we propose a training-free method based on an unbiased estimator of class covariance matrices which only uses first-order statistics in the form of class means communicated by clients to the server. We show how these estimated class covariances can be used to initialize the global classifier, thus exploiting the covariances without actually sharing them. We also show that using only within-class covariances results in a better classifier initialization. Our approach improves performance in the range of 4-26% with exactly the same communication cost when compared to methods sharing only class means and achieves performance competitive or superior to methods sharing second-order statistics with dramatically less communication overhead. The proposed method is much more communication-efficient than federated prompt-tuning methods and still outperforms them. Finally, using our method to initialize classifiers and then performing federated fine-tuning or linear probing again yields better performance. Code is available at https://github.com/dipamgoswami/FedCOF.

cs.LG↗

arXiv2025

Self-Improvement in Multimodal Large Language Models: A Survey

Recent advancements in self-improvement for Large Language Models (LLMs) have efficiently enhanced model capabilities without significantly increasing costs, particularly in terms of human effort. While this area is still relatively young, its extension to the multimodal domain holds immense potential for leveraging diverse data sources and developing more general self-improving models. This survey is the first to provide a comprehensive overview of self-improvement in Multimodal LLMs (MLLMs). We provide a structured overview of the current literature and discuss methods from three perspectives: 1) data collection, 2) data organization, and 3) model optimization, to facilitate the further development of self-improvement in MLLMs. We also include commonly used evaluations and downstream applications. Finally, we conclude by outlining open challenges and future research directions.

cs.CL↗

arXiv2025

Joint Deblurring and 3D Reconstruction for Macrophotography

Macro lens has the advantages of high resolution and large magnification, and 3D modeling of small and detailed objects can provide richer information. However, defocus blur in macrophotography is a long-standing problem that heavily hinders the clear imaging of the captured objects and high-quality 3D reconstruction of them. Traditional image deblurring methods require a large number of images and annotations, and there is currently no multi-view 3D reconstruction method for macrophotography. In this work, we propose a joint deblurring and 3D reconstruction method for macrophotography. Starting from multi-view blurry images captured, we jointly optimize the clear 3D model of the object and the defocus blur kernel of each pixel. The entire framework adopts a differentiable rendering method to self-supervise the optimization of the 3D model and the defocus blur kernel. Extensive experiments show that from a small number of multi-view images, our proposed method can not only achieve high-quality image deblurring but also recover high-fidelity 3D appearance.

cs.CV↗

arXiv2025

On-the-Fly Data Augmentation via Gradient-Guided and Sample-Aware Influence Estimation

Data augmentation has been widely employed to improve the generalization of deep neural networks. Most existing methods apply fixed or random transformations. However, we find that sample difficulty evolves along with the model's generalization capabilities in dynamic training environments. As a result, applying uniform or stochastic augmentations, without accounting for such dynamics, can lead to a mismatch between augmented data and the model's evolving training needs, ultimately degrading training effectiveness. To address this, we introduce SADA, a Sample-Aware Dynamic Augmentation that performs on-the-fly adjustment of augmentation strengths based on each sample's evolving influence on model optimization. Specifically, we estimate each sample's influence by projecting its gradient onto the accumulated model update direction and computing the temporal variance within a local training window. Samples with low variance, indicating stable and consistent influence, are augmented more strongly to emphasize diversity, while unstable samples receive milder transformations to preserve semantic fidelity and stabilize learning. Our method is lightweight, which does not require auxiliary models or policy tuning. It can be seamlessly integrated into existing training pipelines as a plug-and-play module. Experiments across various benchmark datasets and model architectures show consistent improvements of SADA, including +7.3\% on fine-grained tasks and +4.3\% on long-tailed datasets, highlighting the method's effectiveness and practicality.

cs.LG↗

arXiv2025

IMG: Calibrating Diffusion Models via Implicit Multimodal Guidance

Ensuring precise multimodal alignment between diffusion-generated images and input prompts has been a long-standing challenge. Earlier works finetune diffusion weight using high-quality preference data, which tends to be limited and difficult to scale up. Recent editing-based methods further refine local regions of generated images but may compromise overall image quality. In this work, we propose Implicit Multimodal Guidance (IMG), a novel re-generation-based multimodal alignment framework that requires no extra data or editing operations. Specifically, given a generated image and its prompt, IMG a) utilizes a multimodal large language model (MLLM) to identify misalignments; b) introduces an Implicit Aligner that manipulates diffusion conditioning features to reduce misalignments and enable re-generation; and c) formulates the re-alignment goal into a trainable objective, namely Iteratively Updated Preference Objective. Extensive qualitative and quantitative evaluations on SDXL, SDXL-DPO, and FLUX show that IMG outperforms existing alignment methods. Furthermore, IMG acts as a flexible plug-and-play adapter, seamlessly enhancing prior finetuning-based alignment methods. Our code will be available at https://github.com/SHI-Labs/IMG-Multimodal-Diffusion-Alignment.

cs.CV↗

arXiv2025

StarTrail: Concentric Ring Sequence Parallelism for Efficient Near-Infinite-Context Transformer Model Training

Training Transformer models on long sequences in a distributed setting poses significant challenges in terms of efficiency and scalability. Current methods are either constrained by the number of attention heads or excessive communication overheads. To address this problem, we propose StarTrail, a multi-dimensional concentric distributed training system for long sequences, fostering an efficient communication paradigm and providing additional tuning flexibility for communication arrangements. Specifically, StarTrail introduces an extra parallel dimension and divides the peer-to-peer communication into sub-rings to substantially reduce communication volume and avoid bandwidth bottlenecks. Through comprehensive experiments across diverse hardware environments and on both Natural Language Processing (NLP) and Computer Vision (CV) tasks, we demonstrate that our approach significantly surpasses state-of-the-art methods that support Long sequence lengths, achieving performance improvements of up to 77.12% on GPT-style models and up to 114.33% on DiT (Diffusion Transformer) models without affecting the computations results.

cs.DC↗

arXiv2025

A defect-chemistry-informed phase-field model of grain growth in oxide ceramics: application to Fe-doped SrTiO3

Dopants can significantly affect the properties of oxide ceramics through their impact on the property-determined microstructure characteristics such as grain boundary (GB) segregation, space charge layer formation in the GB vicinity, and the grain growth deviating from normal patterns. To support the rational design of oxide ceramics, we propose a defect-chemistry-informed phase-field grain growth model to simulate the microstructure evolution of oxide ceramics. It fully respects the defect-chemistry theory by accounting for the distinct segregation energies and available site densities of charged point defects (oxygen vacancies and acceptor dopants) in both the grain interior and boundaries, and it considers the competing kinetics of defect diffusion and GB movement. The proposed phase-field model is benchmarked against well-known bicrystal models, including the Mott-Schottky and Gouy-Chapman models. Various simulation results are presented to reveal the effect of different defect-chemistry parameters on the space charge layer formation and key microstructural aspects. In particular, simulation results confirm that the solute drag effect alone can lead to skewed grain size distribution that do not follow the log-normal distribution, without any contribution from grain misorientation and other anisotropy. Interestingly, simulations also demonstrate that grain boundary potentials can vary substantially: GBs of larger grains tend to have lower potentials than those of smaller grains. Such heterogeneous GB potential distribution may inspire a new material optimization strategy through microstructure design. This study provides a comprehensive framework for defect-chemistry-consistent investigations of microstructure evolution in polycrystalline oxide ceramics, offering fundamental insights into microscopic processes during critical manufacturing stages.

cond-mat.mtrl-sci↗

arXiv2025

RAPID^3: Tri-Level Reinforced Acceleration Policies for Diffusion Transformer

Diffusion Transformers (DiTs) excel at visual generation yet remain hampered by slow sampling. Existing training-free accelerators - step reduction, feature caching, and sparse attention - enhance inference speed but typically rely on a uniform heuristic or a manually designed adaptive strategy for all images, leaving quality on the table. Alternatively, dynamic neural networks offer per-image adaptive acceleration, but their high fine-tuning costs limit broader applicability. To address these limitations, we introduce RAPID3: Tri-Level Reinforced Acceleration Policies for Diffusion Transformers, a framework that delivers image-wise acceleration with zero updates to the base generator. Specifically, three lightweight policy heads - Step-Skip, Cache-Reuse, and Sparse-Attention - observe the current denoising state and independently decide their corresponding speed-up at each timestep. All policy parameters are trained online via Group Relative Policy Optimization (GRPO) while the generator remains frozen. Meanwhile, an adversarially learned discriminator augments the reward signal, discouraging reward hacking by boosting returns only when generated samples stay close to the original model's distribution. Across state-of-the-art DiT backbones, including Stable Diffusion 3 and FLUX, RAPID3 achieves nearly 3x faster sampling with competitive generation quality.

cs.CV↗

arXiv2025

Fuzzy Reasoning Chain (FRC): An Innovative Reasoning Framework from Fuzziness to Clarity

With the rapid advancement of large language models (LLMs), natural language processing (NLP) has achieved remarkable progress. Nonetheless, significant challenges remain in handling texts with ambiguity, polysemy, or uncertainty. We introduce the Fuzzy Reasoning Chain (FRC) framework, which integrates LLM semantic priors with continuous fuzzy membership degrees, creating an explicit interaction between probability-based reasoning and fuzzy membership reasoning. This transition allows ambiguous inputs to be gradually transformed into clear and interpretable decisions while capturing conflicting or uncertain signals that traditional probability-based methods cannot. We validate FRC on sentiment analysis tasks, where both theoretical analysis and empirical results show that it ensures stable reasoning and facilitates knowledge transfer across different model scales. These findings indicate that FRC provides a general mechanism for managing subtle and ambiguous expressions with improved interpretability and robustness.

cs.CL↗

arXiv2025

Manipulating terahertz phonon-polariton in the ultrastrong coupling regime with bound states in the continuum

The strong coupling between photons and phonons in polar materials gives rise to phonon-polaritons that encapsulate a wealth of physical information, offering crucial tools for the ultrafast terahertz sources and the topological engineering of terahertz light. However, it is still quite challenging to form and manipulate the terahertz phonon-polaritons under the ultrastrong coupling regime till now. In this work, we demonstrate the ultrastrong coupling between the phonon (at 0.95 THz) in a MaPbI 3 film and the metallic bound states in the continuum (BICs) in Au metasurfaces. The Rabi splitting can be continuously tuned from 28% to 48.4% of the phonon frequency by adjusting the parameters (size, shape and period) of Au metasurfaces, reaching the ultrastrong coupling regime. By introducing wavelet transform, the mode evolution information of the terahertz phonon-polariton is successfully extracted. It indicates that the phonon radiation intensity of the MaPbI 3 film is enhanced as the coupling strength is increased. This work not only establishes a new platform for terahertz devices but also opens new avenues for exploring the intricate dynamics of terahertz phonon-polaritons.

physics.optics↗

arXiv2025

DD-Ranking: Rethinking the Evaluation of Dataset Distillation

In recent years, dataset distillation has provided a reliable solution for data compression, where models trained on the resulting smaller synthetic datasets achieve performance comparable to those trained on the original datasets. To further improve the performance of synthetic datasets, various training pipelines and optimization objectives have been proposed, greatly advancing the field of dataset distillation. Recent decoupled dataset distillation methods introduce soft labels and stronger data augmentation during the post-evaluation phase and scale dataset distillation up to larger datasets (e.g., ImageNet-1K). However, this raises a question: Is accuracy still a reliable metric to fairly evaluate dataset distillation methods? Our empirical findings suggest that the performance improvements of these methods often stem from additional techniques rather than the inherent quality of the images themselves, with even randomly sampled images achieving superior results. Such misaligned evaluation settings severely hinder the development of DD. Therefore, we propose DD-Ranking, a unified evaluation framework, along with new general evaluation metrics to uncover the true performance improvements achieved by different methods. By refocusing on the actual information enhancement of distilled datasets, DD-Ranking provides a more comprehensive and fair evaluation standard for future research advancements.

cs.CV↗