arXiv subjects

Kai Wang

Publications and source records attributed to Kai Wang.

At least 343 records · Page 19Linked to original sources

arXiv2025

Landmark-Free Preoperative-to-Intraoperative Registration in Laparoscopic Liver Resection

Liver registration by overlaying preoperative 3D models onto intraoperative 2D frames can assist surgeons in perceiving the spatial anatomy of the liver clearly for a higher surgical success rate. Existing registration methods rely heavily on anatomical landmark-based workflows, which encounter two major limitations: 1) ambiguous landmark definitions fail to provide efficient markers for registration; 2) insufficient integration of intraoperative liver visual information in shape deformation modeling. To address these challenges, in this paper, we propose a landmark-free preoperative-to-intraoperative registration framework utilizing effective self-supervised learning, termed \ourmodel. This framework transforms the conventional 3D-2D workflow into a 3D-3D registration pipeline, which is then decoupled into rigid and non-rigid registration subtasks. \ourmodel~first introduces a feature-disentangled transformer to learn robust correspondences for recovering rigid transformations. Further, a structure-regularized deformation network is designed to adjust the preoperative model to align with the intraoperative liver surface. This network captures structural correlations through geometry similarity modeling in a low-rank transformer network. To facilitate the validation of the registration performance, we also construct an in-vivo registration dataset containing liver resection videos of 21 patients, called \emph{P2I-LReg}, which contains 346 keyframes that provide a global view of the liver together with liver mask annotations and calibrated camera intrinsic parameters. Extensive experiments and user studies on both synthetic and in-vivo datasets demonstrate the superiority and potential clinical applicability of our method.

cs.CV↗

arXiv2025

First-Order Methods for Linearly Constrained Bilevel Optimization

Algorithms for bilevel optimization often encounter Hessian computations, which are prohibitive in high dimensions. While recent works offer first-order methods for unconstrained bilevel problems, the constrained setting remains relatively underexplored. We present first-order linearly constrained optimization methods with finite-time hypergradient stationarity guarantees. For linear equality constraints, we attain $ε$-stationarity in $\widetilde{O}(ε^{-2})$ gradient oracle calls, which is nearly-optimal. For linear inequality constraints, we attain $(δ,ε)$-Goldstein stationarity in $\widetilde{O}(d{δ^{-1} ε^{-3}})$ gradient oracle calls, where $d$ is the upper-level dimension. Finally, we obtain for the linear inequality setting dimension-free rates of $\widetilde{O}({δ^{-1} ε^{-4}})$ oracle complexity under the additional assumption of oracle access to the optimal dual variable. Along the way, we develop new nonsmooth nonconvex optimization methods with inexact oracles. We verify these guarantees with preliminary numerical experiments.

math.OC↗

arXiv2025

Gaussian limit for Pfaffian point processes

We prove a central limit theorem for linear statistics of a broad class of Pfaffian point processes. As an application, we derive Gaussian limits for scaled linear statistics of step functions in the Pfaffian $\mathrm{Sine_4}$ and $\mathrm{Sine}_1$ processes.

math.PR↗

arXiv2025

Faster Vision Mamba is Rebuilt in Minutes via Merged Token Re-training

Vision Mamba has shown close to state of the art performance on computer vision tasks, drawing much interest in increasing it's efficiency. A promising approach is token reduction (that has been successfully implemented in ViTs). Pruning informative tokens in Mamba leads to a high loss of key knowledge and degraded performance. An alternative, of merging tokens preserves more information than pruning, also suffers for large compression ratios. Our key insight is that a quick round of retraining after token merging yeilds robust results across various compression ratios. Empirically, pruned Vims only drop up to 0.9% accuracy on ImageNet-1K, recovered by our proposed framework R-MeeTo in our main evaluation. We show how simple and effective the fast recovery can be achieved at minute-level, in particular, a 35.9% accuracy spike over 3 epochs of training on Vim-Ti. Moreover, Vim-Ti/S/B are re-trained within 5/7/17 minutes, and Vim-S only drops 1.3% with 1.2x (up to 1.5x) speed up in inference.

cs.CV↗

arXiv2025

MedCT: A Clinical Terminology Graph for Generative AI Applications in Healthcare

We introduce the world's first clinical terminology for the Chinese healthcare community, namely MedCT, accompanied by a clinical foundation model MedBERT and an entity linking model MedLink. The MedCT system enables standardized and programmable representation of Chinese clinical data, successively stimulating the development of new medicines, treatment pathways, and better patient outcomes for the populous Chinese community. Moreover, the MedCT knowledge graph provides a principled mechanism to minimize the hallucination problem of large language models (LLMs), therefore achieving significant levels of accuracy and safety in LLM-based clinical applications. By leveraging the LLMs' emergent capabilities of generativeness and expressiveness, we were able to rapidly built a production-quality terminology system and deployed to real-world clinical field within three months, while classical terminologies like SNOMED CT have gone through more than twenty years development. Our experiments show that the MedCT system achieves state-of-the-art (SOTA) performance in semantic matching and entity linking tasks, not only for Chinese but also for English. We also conducted a longitudinal field experiment by applying MedCT and LLMs in a representative spectrum of clinical tasks, including electronic health record (EHR) auto-generation and medical document search for diagnostic decision making. Our study shows a multitude of values of MedCT for clinical workflows and patient outcomes, especially in the new genre of clinical LLM applications. We present our approach in sufficient engineering detail, such that implementing a clinical terminology for other non-English societies should be readily reproducible. We openly release our terminology, models and algorithms, along with real-world clinical datasets for the development.

cs.CL↗

arXiv2025

ORAL: Prompting Your Large-Scale LoRAs via Conditional Recurrent Diffusion

Parameter generation has emerged as a novel paradigm for neural network development, offering an alternative to traditional neural network training by synthesizing high-quality model weights directly. In the context of Low-Rank Adaptation (LoRA) for evolving ($\textit{i.e.}$, constantly updated) large language models (LLMs), this approach promises efficient adaptation without costly retraining. However, existing methods face critical limitations in simultaneously achieving scalability and controllability. In this paper, we introduce $\texttt{ORAL}$, a novel $\textbf{conditional recurrent diffusion}$ framework that addresses these challenges. $\texttt{ORAL}$ incorporates a novel conditioning mechanism that integrates model architecture and textual task specifications, enabling the generation of task-specific LoRA parameters that can seamlessly transfer across evolving foundation models. Our approach successfully scales to billions-of-parameter LLMs and maintains controllability. Through extensive experiments across seven language tasks, four vision tasks, and three multimodal tasks using five pre-trained LLMs, we demonstrate that $\texttt{ORAL}$ generates high-quality LoRA parameters that achieve comparable or superior performance to vanilla trained counterparts.

cs.LG↗

arXiv2025

MDK12-Bench: A Multi-Discipline Benchmark for Evaluating Reasoning in Multimodal Large Language Models

Multimodal reasoning, which integrates language and visual cues into problem solving and decision making, is a fundamental aspect of human intelligence and a crucial step toward artificial general intelligence. However, the evaluation of multimodal reasoning capabilities in Multimodal Large Language Models (MLLMs) remains inadequate. Most existing reasoning benchmarks are constrained by limited data size, narrow domain coverage, and unstructured knowledge distribution. To close these gaps, we introduce MDK12-Bench, a multi-disciplinary benchmark assessing the reasoning capabilities of MLLMs via real-world K-12 examinations. Spanning six disciplines (math, physics, chemistry, biology, geography, and information science), our benchmark comprises 140K reasoning instances across diverse difficulty levels from primary school to 12th grade. It features 6,827 instance-level knowledge point annotations based on a well-organized knowledge structure, detailed answer explanations, difficulty labels and cross-year partitions, providing a robust platform for comprehensive evaluation. Additionally, we present a novel dynamic evaluation framework to mitigate data contamination issues by bootstrapping question forms, question types, and image styles during evaluation. Extensive experiment on MDK12-Bench reveals the significant limitation of current MLLMs in multimodal reasoning. The findings on our benchmark provide insights into the development of the next-generation models. Our data and codes are available at https://github.com/LanceZPF/MDK12.

cs.CV↗

arXiv2025

KERMIT: Knowledge Graph Completion of Enhanced Relation Modeling with Inverse Transformation

Knowledge graph completion (KGC) revolves around populating missing triples in a knowledge graph using available information. Text-based methods, which depend on textual descriptions of triples, often encounter difficulties when these descriptions lack sufficient information for accurate prediction-an issue inherent to the datasets and not easily resolved through modeling alone. To address this and ensure data consistency, we first use large language models (LLMs) to generate coherent descriptions, bridging the semantic gap between queries and answers. Secondly, we utilize inverse relations to create a symmetric graph, thereby providing augmented training samples for KGC. Additionally, we employ the label information inherent in knowledge graphs (KGs) to enhance the existing contrastive framework, making it fully supervised. These efforts have led to significant performance improvements on the WN18RR and FB15k-237 datasets. According to standard evaluation metrics, our approach achieves a 4.2% improvement in Hit@1 on WN18RR and a 3.4% improvement in Hit@3 on FB15k-237, demonstrating superior performance.

cs.CL↗

arXiv2025

Dynamic Vision Mamba

Mamba-based vision models have gained extensive attention as a result of being computationally more efficient than attention-based models. However, spatial redundancy still exists in these models, represented by token and block redundancy. For token redundancy, we analytically find that early token pruning methods will result in inconsistency between training and inference or introduce extra computation for inference. Therefore, we customize token pruning to fit the Mamba structure by rearranging the pruned sequence before feeding it into the next Mamba block. For block redundancy, we allow each image to select SSM blocks dynamically based on an empirical observation that the inference speed of Mamba-based vision models is largely affected by the number of SSM blocks. Our proposed method, Dynamic Vision Mamba (DyVM), effectively reduces FLOPs with minor performance drops. We achieve a reduction of 35.2\% FLOPs with only a loss of accuracy of 1.7\% on Vim-S. It also generalizes well across different Mamba vision model architectures and different vision tasks. Our code will be made public.

cs.CV↗

arXiv2025

Efficient Computation of Hyper-triangles on Hypergraphs

Hypergraphs, which use hyperedges to capture groupwise interactions among different entities, have gained increasing attention recently for their versatility in effectively modeling real-world networks. In this paper, we study the problem of computing hyper-triangles (formed by three fully-connected hyperedges), which is a basic structural unit in hypergraphs. Although existing approaches can be adopted to compute hyper-triangles by exhaustively examining hyperedge combinations, they overlook the structural characteristics distinguishing different hyper-triangle patterns. Consequently, these approaches lack specificity in computing particular hyper-triangle patterns and exhibit low efficiency. In this paper, we unveil a new formation pathway for hyper-triangles, transitioning from hyperedges to hyperwedges before assembling into hyper-triangles, and classify hyper-triangle patterns based on hyperwedges. Leveraging this insight, we introduce a two-step framework to reduce the redundant checking of hyperedge combinations. Under this framework, we propose efficient algorithms for computing a specific pattern of hyper-triangles. Approximate algorithms are also devised to support estimated counting scenarios. Furthermore, we introduce a fine-grained hypergraph clustering coefficient measurement that can reflect diverse properties of hypergraphs based on different hyper-triangle patterns. Extensive experimental evaluations conducted on 11 real-world datasets validate the effectiveness and efficiency of our proposed techniques.

cs.DS↗

arXiv2025

Slow-Fast Architecture for Video Multi-Modal Large Language Models

Balancing temporal resolution and spatial detail under limited compute budget remains a key challenge for video-based multi-modal large language models (MLLMs). Existing methods typically compress video representations using predefined rules before feeding them into the LLM, resulting in irreversible information loss and often ignoring input instructions. To address this, we propose a novel slow-fast architecture that naturally circumvents this trade-off, enabling the use of more input frames while preserving spatial details. Inspired by how humans first skim a video before focusing on relevant parts, our slow-fast design employs a dual-token strategy: 1) "fast" visual tokens -- a compact set of compressed video features -- are fed into the LLM alongside text embeddings to provide a quick overview; 2) "slow" visual tokens -- uncompressed video features -- are cross-attended by text embeddings through specially designed hybrid decoder layers, enabling instruction-aware extraction of relevant visual details with linear complexity. We conduct systematic exploration to optimize both the overall architecture and key components. Experiments show that our model significantly outperforms self-attention-only baselines, extending the input capacity from 16 to 128 frames with just a 3% increase in computation, and achieving a 16% average performance improvement across five video understanding benchmarks. Our 7B model achieves state-of-the-art performance among models of similar size. Furthermore, our slow-fast architecture is a plug-and-play design that can be integrated into other video MLLMs to improve efficiency and scalability.

cs.CV↗

arXiv2025

Emphasizing Discriminative Features for Dataset Distillation in Complex Scenarios

Dataset distillation has demonstrated strong performance on simple datasets like CIFAR, MNIST, and TinyImageNet but struggles to achieve similar results in more complex scenarios. In this paper, we propose EDF (emphasizes the discriminative features), a dataset distillation method that enhances key discriminative regions in synthetic images using Grad-CAM activation maps. Our approach is inspired by a key observation: in simple datasets, high-activation areas typically occupy most of the image, whereas in complex scenarios, the size of these areas is much smaller. Unlike previous methods that treat all pixels equally when synthesizing images, EDF uses Grad-CAM activation maps to enhance high-activation areas. From a supervision perspective, we downplay supervision signals that have lower losses, as they contain common patterns. Additionally, to help the DD community better explore complex scenarios, we build the Complex Dataset Distillation (Comp-DD) benchmark by meticulously selecting sixteen subsets, eight easy and eight hard, from ImageNet-1K. In particular, EDF consistently outperforms SOTA results in complex scenarios, such as ImageNet-1K subsets. Hopefully, more researchers will be inspired and encouraged to improve the practicality and efficacy of DD. Our code and benchmark will be made public at https://github.com/NUS-HPC-AI-Lab/EDF.

cs.CV↗

arXiv2025

Deformations of Nijenhuis Lie algebras and Nijenhuis Lie algebroids

This paper is the second in a series dedicated to the operadic study of Nijenhuis structures, focusing on Nijenhuis Lie algebras and Nijenhuis geometry. We introduce the concept of homotopy Nijenhuis Lie algebras and establish that the differential graded (=dg) operad $\mathfrak{NjL}_{\infty}$ governing these structures serves as the minimal model of the operad $\mathfrak{NjL}$ for Nijenhuis Lie algebras. We construct an $L_\infty$-algebra that encodes the simultaneous deformations of Lie brackets and Nijenhuis operators, leading to the deformation cochain complex and an associated cohomology theory for Nijenhuis Lie algebras. Extending these ideas to geometry, we investigate the deformations of geometric Nijenhuis structures. We introduce the notion of a Nijenhuis Lie algebroid-a Lie algebroid equipped with a Nijenhuis structure, which generalizes the classical Nijenhuis structure on vector fields of manifolds. Using the framework of dg manifolds, we construct an $L_\infty$-algebra that governs the simultaneous deformations of Lie algebroid structures and Nijenhuis operators. As a computational application, we prove that a certain class of Nijenhuis operators satisfies the Poincaré Lemma, meaning its cohomology vanishes, which confirms a conjecture by Bolsinov and Konyaev.

math.DG↗

arXiv2025

Advancing CAN Network Security through RBM-Based Synthetic Attack Data Generation for Intrusion Detection Systems

The rapid development of network technologies and industrial intelligence has augmented the connectivity and intelligence within the automotive industry. Notably, in the Internet of Vehicles (IoV), the Controller Area Network (CAN), which is crucial for the communication of electronic control units but lacks inbuilt security measures, has become extremely vulnerable to severe cybersecurity threats. Meanwhile, the efficacy of Intrusion Detection Systems (IDS) is hampered by the scarcity of sufficient attack data for robust model training. To overcome this limitation, we introduce a novel methodology leveraging the Restricted Boltzmann Machine (RBM) to generate synthetic CAN attack data, thereby producing training datasets with a more balanced sample distribution. Specifically, we design a CAN Data Processing Module for transforming raw CAN data into an RBM-trainable format, and a Negative Sample Generation Module to generate data reflecting the distribution of CAN data frames denoting network intrusions. Experimental results show the generated data significantly improves IDS performance, with CANet accuracy rising from 0.6477 to 0.9725 and EfficientNet from 0.1067 to 0.1555. Code is available at https://github.com/wangkai-tech23/CANDataSynthetic.

cs.CR↗

arXiv2025

A Closer Look at Time Steps is Worthy of Triple Speed-Up for Diffusion Model Training

Training diffusion models is always a computation-intensive task. In this paper, we introduce a novel speed-up method for diffusion model training, called, which is based on a closer look at time steps. Our key findings are: i) Time steps can be empirically divided into acceleration, deceleration, and convergence areas based on the process increment. ii) These time steps are imbalanced, with many concentrated in the convergence area. iii) The concentrated steps provide limited benefits for diffusion training. To address this, we design an asymmetric sampling strategy that reduces the frequency of steps from the convergence area while increasing the sampling probability for steps from other areas. Additionally, we propose a weighting strategy to emphasize the importance of time steps with rapid-change process increments. As a plug-and-play and architecture-agnostic approach, SpeeD consistently achieves 3-times acceleration across various diffusion architectures, datasets, and tasks. Notably, due to its simple design, our approach significantly reduces the cost of diffusion model training with minimal overhead. Our research enables more researchers to train diffusion models at a lower cost.

cs.LG↗

arXiv2025

Motif Counting in Complex Networks: A Comprehensive Survey

Motif counting plays a crucial role in understanding the structural properties of networks. By computing motif frequencies, researchers can draw key insights into the structural properties of the underlying network. As networks become increasingly complex, different graph models have been proposed, giving rise to diverse motif patterns. These variations introduce unique computational challenges that require specialized algorithms tailored to specific motifs within different graph structures. This survey provides a comprehensive and structured overview of motif counting techniques across general graphs, heterogeneous graphs, and hypergraphs. We categorize existing algorithms according to their underlying computational strategies, emphasizing key similarities and distinctions. In addition to reviewing current methodologies, we examine their strengths, limitations, and computational trade-offs. Furthermore, we explore future directions in motif counting, including scalable implementations to improve efficiency in large-scale networks, algorithmic adaptations for dynamic, temporal, and attributed graphs, and deeper integration with large language models (LLMs) and graph-based retrieval-augmented generation (GraphRAG). By offering a detailed analysis of these approaches, this survey aims to support researchers and practitioners in advancing motif counting for increasingly complex network data.

cs.SI↗

arXiv2025

STATGRAPH: Effective In-vehicle Intrusion Detection via Multi-view Statistical Graph Learning

In-vehicle network (IVN) is facing complex external cyber-attacks, especially the emerging masquerade attacks with extremely high difficulty of detection while serious damaging effects. In this paper, we propose the STATGRAPH, which is an effective and fine-grained intrusion detection methodology for IVN security services via multi-view statistical graph learning on in-vehicle controller area network (CAN) messages with insight into their variations in periodicity, payload and signal combinations. Specifically, STATGRAPH generates two statistical graphs, timing correlation graph (TCG) and coupling relationship graph (CRG), in every CAN message detection window, where edge attributes in TCGs represent temporal correlation between different message IDs while edge attributes in CRGs denote the neighbour relationship and contextual similarity. Besides, a lightweight shallow layered graph convolution network is trained based on graph property of TCGs and CRGs, which learns the universal laws of various patterns more effectively and further enhance the performance of detection. To address the problem of insufficient attack types in previous intrusion detection, we select two real in-vehicle CAN datasets covering five new instances of sophisticated and stealthy masquerade attacks that are never investigated before. Experimental result shows STATGRAPH improves both detection granularity and detection performance over state-of-the-art intrusion detection methods. Code is available at https://github.com/wangkai-tech23/StatGraph.

cs.NI↗

arXiv2025

ATOM: Attention Mixer for Efficient Dataset Distillation

Recent works in dataset distillation seek to minimize training expenses by generating a condensed synthetic dataset that encapsulates the information present in a larger real dataset. These approaches ultimately aim to attain test accuracy levels akin to those achieved by models trained on the entirety of the original dataset. Previous studies in feature and distribution matching have achieved significant results without incurring the costs of bi-level optimization in the distillation process. Despite their convincing efficiency, many of these methods suffer from marginal downstream performance improvements, limited distillation of contextual information, and subpar cross-architecture generalization. To address these challenges in dataset distillation, we propose the ATtentiOn Mixer (ATOM) module to efficiently distill large datasets using a mixture of channel and spatial-wise attention in the feature matching process. Spatial-wise attention helps guide the learning process based on consistent localization of classes in their respective images, allowing for distillation from a broader receptive field. Meanwhile, channel-wise attention captures the contextual information associated with the class itself, thus making the synthetic image more informative for training. By integrating both types of attention, our ATOM module demonstrates superior performance across various computer vision datasets, including CIFAR10/100 and TinyImagenet. Notably, our method significantly improves performance in scenarios with a low number of images per class, thereby enhancing its potential. Furthermore, we maintain the improvement in cross-architectures and applications such as neural architecture search.

cs.CV↗