Search arXiv⌕ Search

arXiv subjects

Kai Wang

Publications and source records attributed to Kai Wang.

At least 181 records · Page 10Linked to original sources
arXiv2026

Mixture of Heterogeneous Grouped Experts for Language Modeling

Large Language Models (LLMs) based on Mixture-of-Experts (MoE) are pivotal in industrial applications for their ability to scale performance efficiently. However, standard MoEs enforce uniform expert sizes,creating a rigidity that fails to align computational costs with varying token-level complexity. While heterogeneous expert architectures attempt to address this by diversifying expert sizes, they often suffer from significant system-level challenges, specifically unbalanced GPU utilization and inefficient parameter utilization, which hinder practical deployment. To bridge the gap between theoretical heterogeneity and robust industrial application, we propose Mixture of Heterogeneous Grouped Experts (MoHGE) which introduces a two-level routing mechanism to enable flexible, resource-aware expert combinations. To optimize inference efficiency, we propose a Group-Wise Auxiliary Loss, which dynamically steers tokens to the most parameter-efficient expert groups based on task difficulty. To address the critical deployment challenge of GPU load balancing, we introduce an All-size Group-decoupling Allocation strategy coupled with an Intra-Group Experts Auxiliary Loss. These mechanisms collectively ensure uniform computation distribution across GPUs. Extensive evaluations demonstrate that MoHGE matches the performance of MoE architectures while reducing the total parameters by approximately 20% and maintaining balanced GPU utilization. Our work establishes a scalable paradigm for resource-efficient MoE design, offering a practical solution for optimizing inference costs in real-world scenarios. The code is publicly available at https://github.com/UnicomAI/MoHGE.

cs.CL↗
arXiv2026

GLM-5 Serving Parameter Tuning for OpenClaw: Single-Deployment MaaS Inference Optimization for Long-Context Agent Workloads

OpenClaw requests are dominated by long, tool-augmented prefixes, including system prompts, conversation history, and tool outputs fed back into the context window. For this workload, with about 28k-30k input tokens and 500 output tokens per request, serving quality is governed by throughput, TTFT, and tail latency rather than short-prompt throughput alone. This report studies GLM-5 serving-parameter tuning within a MaaS multi-model inference optimization architecture. The scope is the Single-Node Optimization block of the inference-optimization layer, where chunked prefill, tensor parallelism (TP), pipeline parallelism (PP), and request concurrency are tuned for one GLM-5 serving deployment; in this report, "Single-Node Optimization" denotes the architecture block, while experiments run on a two-node, sixteen-GPU cluster. Within the tested space, the best configuration is chunked-prefill-size=3072, tp=4, pp-size=4, and max-running-requests=24. Compared with the conservative 2048/4/4/16 baseline, it increases request throughput from 0.43 to 0.48 req/s and total token throughput from 9029.64 to 9993.23 tok/s, while reducing average TTFT from 8.98 to 6.69 s and latency P90 from 40.23 to 32.64 s. Under the same hardware footprint, this corresponds to an estimated 10.4% lower serving cost per request and 9.6% lower cost per token. The results show that the optimum is workload-specific: larger chunk sizes and deeper queueing do not monotonically improve performance. We therefore recommend 3072 / tp4 / pp4 / max24 as the default OpenClaw deployment profile.

cs.DC↗
arXiv2026

Counting Butterflies over Streaming Bipartite Graphs with Duplicate Edges

Bipartite graphs are commonly used to model relationships between two distinct entities in real-world applications, such as user-product interactions, user-movie ratings and collaborations between authors and publications. A butterfly (a 2x2 bi-clique) is a critical substructure in bipartite graphs, playing a significant role in tasks like community detection, fraud detection, and link prediction. As more real-world data is presented in a streaming format, efficiently counting butterflies in streaming bipartite graphs has become increasingly important. However, most existing algorithms typically assume that duplicate edges are absent, which is hard to hold in real-world graph streams, as a result, they tend to sample edges that appear multiple times, leading to inaccurate results. The only algorithm designed to handle duplicate edges is FABLE, but it suffers from significant limitations, including high variance, substantial time complexity, and memory inefficiency due to its reliance on a priority queue. To overcome these limitations, we introduce DEABC (Duplicate-Edge-Aware Butterfly Counting), an innovative method that uses bucket-based priority sampling to accurately estimate the number of butterflies, accounting for duplicate edges. Compared to existing methods, DEABC significantly reduces memory usage by storing only the essential sampled edge data while maintaining high accuracy. We provide rigorous proofs of the unbiasedness and variance bounds for DEABC, ensuring they achieve high accuracy. We compare DEABC with state-of-the-art algorithms on real-world streaming bipartite graphs. The results show that our DEABC outperforms existing methods in memory efficiency and accuracy, while also achieving significantly higher throughput.

cs.DS↗
arXiv2026

KAConvNet: Kolmogorov-Arnold Convolutional Networks for Vision Recognition

The Convolutional Neural Networks (CNNs) have been the dominant and effective approach for general computer vision tasks. Recently, Kolmogorov-Arnold neural networks (KANs), based on the Kolmogorov-Arnold representation theorem, have shown potential to replace Multi-Layer Perceptrons (MLPs) in deep learning. KANs, which use learnable nonlinear activations on edges and simple summation on nodes, offer fewer parameters and greater explainability compared to MLPs. However, there has been limited exploration of integrating the Kolmogorov-Arnold representation theorem with convolutional methods for computer vision tasks. Existing attempts have merely replaced learnable activation functions with weights, undermining KANs' theoretical foundation and limiting their potential effectiveness. Additionally, the B-spline curves used in KANs suffer from computational inefficiency and a tendency to overfit. In this paper, we propose a novel Kolmogorov-Arnold Convolutional Layer that deeply integrates the Kolmogorov-Arnold representation theorem with convolution. This layer provides stronger method interpretability because it is based on established mathematical theorems and its design has theoretical alignment. Building on the Kolmogorov-Arnold Convolutional Layer, we design an efficient network architecture called KAConvNet, which outperforms existing methods combining KAN and convolution, and achieves competitive performance compared to mainstream ViTs and CNNs. We believe that our work offers valuable insight into the field of artificial intelligence and will inspire the development of more innovative CNNs in the 2020s. The code is publicly available at https://github.com/UnicomAI/KAConvNet.

cs.CV↗
arXiv2026

Accurate Nanoscale Mapping of Electric Fields across Random Grain Boundaries in Polycrystalline Oxides Using Precession-Assisted 4D-STEM

Space charge layers (SCLs) at grain boundaries play a crucial role in modulating local electric fields and influencing the functional properties of materials, such as oxygen vacancy migration and ionic conductivity in oxide ceramics. However, the direct experimental analysis of such localized electric fields and the corresponding charge distribution remains challenging. Conventional center-of-mass (CoM) analysis in scanning transmission electron microscopy differential phase contrast (STEM-DPC) is strongly affected by orientation-dependent contrast and dynamical scattering. Here, we demonstrate that combining electron beam precession with advanced post-processing, employing iterative edge detection via a Sobel filter and singular value decomposition (SVD), enables reliable and accurate, unbiased diffraction shift measurements with minimal crystallographic artefacts. The new method accurately refines the central disk position in nanobeam electron diffraction (NBED) patterns and thus significantly improves the extraction of the local electric field and corresponding charge distribution. Comparative analysis with conventional CoM methods shows superior accuracy and robustness for random grain boundaries in BaTiO3 and SrTiO3 as exemplary case studies. The experimental work is complemented by atomistic simulations to separate the electric field of the SCL from the mean inner potential difference of the grain boundary and the elemental segregation around the grain boundary. The in-depth analysis shows that our approach enables high-fidelity mapping of electromagnetic fields and their charge distribution in complex polycrystalline specimens, laying the groundwork for improved quantitative analysis using STEM-DPC.

cond-mat.mtrl-sci↗
arXiv2026

From homotopy Rota-Baxter algebras to Pre-Calabi-Yau and homotopy double Poisson algebras

In this paper, we investigate pre-Calabi-Yau algebras and homotopy double Poisson algebras arising from homotopy Rota-Baxter structures. We introduce the notion of cyclic homotopy Rota-Baxter algebras, a class of homotopy Rota-Baxter algebras endowed with additional cyclic symmetry, and present a construction of such structures via a process called cyclic completion. We further introduce the concept of interactive pairs, consisting of two differential graded algebras-designated as the acting algebra and the base algebra-interacting through compatible module structures. We prove that if the acting algebra carries a suitable cyclic homotopy Rota-Baxter structure, then the base algebra inherits a natural pre-Calabi-Yau structure. Using the correspondence established by Fernandez and Herscovich between pre-Calabi-Yau algebras and homotopy double Poisson algebras, we describe the resulting homotopy Poisson structure on the base algebra in terms of homotopy Rota-Baxter algebra structure. In particular, we show that a module over an ultracyclic (resp. cyclic) homotopy Rota-Baxter algebra admits a (resp. cyclic) homotopy double Lie algebra structure.

math.RT↗
arXiv2026

Graph-Guided Adaptive Channel Elimination for KV Cache Compression

Large Language Models have revolutionized natural language processing, achieving unprecedented success across a vast range of tasks. However, their practical application in long-context scenarios is severely hampered by the formidable memory footprint of the Key-Value cache. While channel pruning has emerged as a promising compression strategy, existing methods evaluate channel importance in isolation, fundamentally ignoring the inter-channel interactions that collectively dictate model performance. This oversight leads to suboptimal pruning decisions. To address this, we introduce \textbf{GRACE} (\textbf{GR}aph-guided \textbf{A}daptive \textbf{C}hannel \textbf{E}limination), a novel framework that reframes KV cache compression as a graph-based optimization problem. GRACE models channels as nodes and their interactions as weighted edges, enabling the identification of a near-optimal channel subset for pruning by minimizing the reconstruction error of the attention weight matrix. Furthermore, GRACE incorporates an adaptive protection mechanism that shields salient key channels from removal, ensuring a robust autoregressive decoding process. Extensive experiments show that GRACE can reduce KV cache size by 60\% with negligible performance degradation, consistently outperforming the state-of-the-art method.

eess.SP↗
arXiv2026

GeV gamma-ray emission in the field of the shell-type supernova remnant Vela Jr revisited

We present an updated analysis of the gigaelectronvolt (GeV) gamma-ray emission from the shell-type supernova remnant (SNR) RX J0852.0-4622 (Vela Jr) using 15 yr of Fermi Large Area Telescope (Fermi-LAT) data. We quantitatively model the GeV morphology and find that it is best described by the masked H.E.S.S. shell template, indicating that the embedded pulsar wind nebula (PWN) contributes little to the GeV flux. The 0.1-500 GeV spectrum is well fitted by a hard power law with a photon index of $1.77 \pm 0.03$ and connects smoothly to the teraelectronvolt (TeV) spectrum, confirming previous results with improved precision. We further construct an independent eROSITA shell template and derive the 1-5 keV X-ray spectral energy distribution (SED) of the whole remnant, which provides new constraints on the synchrotron emission. We model the multi-wavelength (MWL) SED with a pure leptonic model and a hybrid lepton-hadron model. While the pure leptonic model reproduces the overall broadband shape, the hybrid model provides a better statistical description of the same dataset, supporting a mixed-origin picture in which the hadronic contribution is mainly relevant in the GeV band and the TeV emission remains predominantly leptonic.

astro-ph.HE↗
arXiv2026

The Salami Slicing Threat: Exploiting Cumulative Risks in LLM Systems

Large Language Models (LLMs) face prominent security risks from jailbreaking, a practice that manipulates models to bypass built-in security constraints and generate unethical or unsafe content. Among various jailbreak techniques, multi-turn jailbreak attacks are more covert and persistent than single-turn counterparts, exposing critical vulnerabilities of LLMs. However, existing multi-turn jailbreak methods suffer from two fundamental limitations that affect the actual impact in real-world scenarios: (a) As models become more context-aware, any explicit harmful trigger is increasingly likely to be flagged and blocked; (b) Successful final-step triggers often require finely tuned, model-specific contexts, making such attacks highly context-dependent. To fill this gap, we propose \textit{Salami Slicing Risk}, which operates by chaining numerous low-risk inputs that individually evade alignment thresholds but cumulatively accumulate harmful intent to ultimately trigger high-risk behaviors, without heavy reliance on pre-designed contextual structures. Building on this risk, we develop Salami Attack, an automatic framework universally applicable to multiple model types and modalities. Rigorous experiments demonstrate its state-of-the-art performance across diverse models and modalities, achieving over 90\% Attack Success Rate on GPT-4o and Gemini, as well as robustness against real-world alignment defenses. We also proposed a defense strategy to constrain the Salami Attack by at least 44.8\% while achieving a maximum blocking rate of 64.8\% against other multi-turn jailbreak attacks. Our findings provide critical insights into the pervasive risks of multi-turn jailbreaking and offer actionable mitigation strategies to enhance LLM security.

cs.CR↗
arXiv2026

Back to Basics: Let Conversational Agents Remember with Just Retrieval and Generation

Existing conversational memory systems rely on complex hierarchical summarization or reinforcement learning to manage long-term dialogue history, yet remain vulnerable to context dilution as conversations grow. In this work, we offer a different perspective: the primary bottleneck may lie not in memory architecture, but in the \textit{Signal Sparsity Effect} within the latent knowledge manifold. Through controlled experiments, we identify two key phenomena: \textit{Decisive Evidence Sparsity}, where relevant signals become increasingly isolated with longer sessions, leading to sharp degradation in aggregation-based methods; and \textit{Dual-Level Redundancy}, where both inter-session interference and intra-session conversational filler introduce large amounts of non-informative content, hindering effective generation. Motivated by these insights, we propose \method, a minimalist framework that brings conversational memory back to basics, relying solely on retrieval and generation via Turn Isolation Retrieval (TIR) and Query-Driven Pruning (QDP). TIR replaces global aggregation with a max-activation strategy to capture turn-level signals, while QDP removes redundant sessions and conversational filler to construct a compact, high-density evidence set. Extensive experiments on multiple benchmarks demonstrate that \method achieves robust performance across diverse settings, consistently outperforming strong baselines while maintaining high efficiency in tokens and latency, establishing a new minimalist baseline for conversational memory.

cs.CL↗
arXiv2026

Adaptive Action Chunking at Inference-time for Vision-Language-Action Models

In Vision-Language-Action (VLA) models, action chunking (i.e., executing a sequence of actions without intermediate replanning) is a key technique to improve robotic manipulation abilities. However, a large chunk size reduces the model's responsiveness to new information, while a small one increases the likelihood of mode-jumping, jerky behavior resulting from discontinuities between chunks. Therefore, selecting the optimal chunk size is an urgent demand to balance the model's reactivity and consistency. Unfortunately, a dominant trend in current VLA models is an empirical fixed chunk length at inference-time, hindering their superiority and scalability across diverse manipulation tasks. To address this issue, we propose a novel Adaptive Action Chunking (AAC) strategy, which exploits action entropy as the cue to adaptively determine the chunk size based on current predictions. Extensive experiments on a wide range of simulated and real-world robotic manipulation tasks have demonstrated that our approach substantially improves performance over the state-of-the-art alternatives. The videos and source code are publicly available at https://lance-lot.github.io/adaptive-chunking.github.io/.

cs.RO↗
arXiv2026

Seyfert Galaxies as Neutrino Sources: An Outflow$-$Cloud Interaction Perspective

Following the identification of the first confirmed individual neutrino source, Seyfert galaxies have emerged as the most prominent class of high-energy neutrino emitters. In this work, we perform a detailed investigation of the outflow--cloud interaction scenario for neutrino production in Seyfert nuclei. In this framework, fast AGN-driven winds collide with clumpy gas clouds in the nuclear region, forming bow shocks that efficiently accelerate cosmic-ray protons. The accelerated protons subsequently interact with cold protons from the outflows via inelastic proton--proton ($pp$) collisions, producing high-energy neutrinos, while the photomeson ($pγ$) process with disk photons may provide a subdominant contribution at the highest energies. Applying this model to five neutrino-associated Seyfert galaxies, we successfully reproduce the observed TeV neutrino fluxes without violating existing gamma-ray constraints. By integrating over the Seyfert population using X-ray luminosity functions, we further demonstrate that Seyfert galaxies can account for a substantial fraction of the diffuse astrophysical neutrino background in the $10^4-10^5~{\rm GeV}$ energy range.

astro-ph.HE↗
arXiv2026

LumiCtrl : Learning Illuminant Prompts for Lighting Control in Personalized Text-to-Image Models

Text-to-image (T2I) models have demonstrated remarkable progress in creative image generation, yet they still lack precise control over scene illuminants which is a crucial factor for content designers to manipulate visual aesthetics of generated images. In this paper, we present an illuminant personalization method named LumiCtrl that learns illuminant prompt given single image of the object. LumiCtrl consists of three components: given an image of the object, our method apply (a) physics-based illuminant augmentation along with Planckian locus to create fine-tuning variants under standard illuminants; (b) Edge-Guided Prompt Disentanglement using frozen ControlNet to ensure prompts focus on illumination, not the structure; and (c) a Masked Reconstruction Loss that focuses learning on foreground object while allowing background to adapt contextually which enables what we call Contextual Light Adaptation. We qualitatively and quantitatively compare LumiCtrl against other T2I customization methods. The results show that LumiCtrl achieves significantly better illuminant fidelity, aesthetic quality, and scene coherence compared to existing baselines. A human preference study further confirms the strong user preference for LumiCtrl generations.

cs.CV↗
arXiv2026

Beyond Loss Values: Robust Dynamic Pruning via Loss Trajectory Alignment

Existing dynamic data pruning methods often fail under noisy-label settings, as they typically rely on per-sample loss as the ranking criterion. This could mistakenly lead to preserving noisy samples due to their high loss values, resulting in significant performance drop. To address this, we propose AlignPrune, a noise-robust module designed to enhance the reliability of dynamic pruning under label noise. Specifically, AlignPrune introduces the Dynamic Alignment Score (DAS), which is a loss-trajectory-based criterion that enables more accurate identification of noisy samples, thereby improving pruning effectiveness. As a simple yet effective plug-and-play module, AlignPrune can be seamlessly integrated into state-of-the-art dynamic pruning frameworks, consistently outperforming them without modifying either the model architecture or the training pipeline. Extensive experiments on five widely-used benchmarks across various noise types and pruning ratios demonstrate the effectiveness of AlignPrune, boosting accuracy by up to 6.3\% over state-of-the-art baselines. Our results offer a generalizable solution for pruning under noisy data, encouraging further exploration of learning in real-world scenarios. Code is available at: https://github.com/leonqin430/AlignPrune.

cs.CV↗
arXiv2026

Impact of charge transition levels on grain boundary properties in acceptor doped oxide ceramics: A phase-field study

Advanced doping strategies enable oxide ceramic functionalities by tailoring bulk defect chemistry and space-charge-layer (SCL) behavior at interfaces. Charge transition levels (CTLs), defined as the Fermi level at which a defect changes its stable charge state, play a central role. Their alignment governs bulk defect chemistry, while their bending within SCLs induces additional charge-state transitions. Incorporating CTLs is therefore essential for a consistent description of defect equilibria and SCL formation. In this work, we propose a defect-chemistry-consistent phase-field model explicitly coupled with CTLs to investigate their role in SCL evolution. The model includes multivalent oxygen vacancies, multivalent acceptor dopants, electrons, and holes. It is applied to Fe-doped SrTiO3 over wide ranges of oxygen partial pressure and temperature, capturing both symmetric SCLs at stationary grain boundaries and asymmetric SCLs during migration. Two distinct grain boundary types, slow and fast boundaries, emerge during migration, consistent with experimental observations. Simulations reveal that CTL-governed bulk defect chemistry, together with CTL-induced charge-state transitions within SCLs, critically determine SCL characteristics. Moreover, CTL-mediated hole transport is significantly faster than acceptor dopant diffusion, modulating solute drag and grain boundary kinetics. Finally, the model predicts grain boundary properties dependent on both thermal history and boundary type, with slow and fast boundaries exhibiting distinct behaviors. This framework links defect chemistry, Fermi level, CTLs, and grain boundary kinetics, providing new insights for designing oxide ceramics with tailored properties.

cond-mat.mtrl-sci↗
arXiv2026

OmniSonic: Towards Universal and Holistic Audio Generation from Video and Text

In this paper, we propose Universal Holistic Audio Generation (UniHAGen), a task for synthesizing comprehensive auditory scenes that include both on-screen and off-screen sounds across diverse domains (e.g., ambient events, musical instruments, and human speech). Prior video-conditioned audio generation models typically focus on producing on-screen environmental sounds that correspond to visible sounding events, neglecting off-screen auditory events. While recent holistic joint text-video-to-audio generation models aim to produce auditory scenes with both on- and off-screen sound but they are limited to non-speech sounds, lacking the ability to generate or integrate human speech. To overcome these limitations, we introduce OmniSonic, a flow-matching-based diffusion framework jointly conditioned on video and text. It features a TriAttn-DiT architecture that performs three cross-attention operations to process on-screen environmental sound, off-screen environmental sound, and speech conditions simultaneously, with a Mixture-of-Experts (MoE) gating mechanism that adaptively balances their contributions during generation. Furthermore, we construct UniHAGen-Bench, a new benchmark with over one thousand samples covering three representative on/off-screen speech-environment scenarios. Extensive experiments show that OmniSonic consistently outperforms state-of-the-art approaches on both objective metrics and human evaluations, establishing a strong baseline for universal and holistic audio generation. Project page: https://weiguopian.github.io/OmniSonic_webpage/

cs.SD↗
arXiv2026

EgoDemoGen: Egocentric Demonstration Generation for Viewpoint Generalization in Robotic Manipulation

Imitation learning based visuomotor policies have achieved strong performance in robotic manipulation, yet they often remain sensitive to egocentric viewpoint shifts. Unlike third-person viewpoint changes that only move the camera, egocentric shifts simultaneously alter both the camera pose and the robot action coordinate frame, making it necessary to jointly transfer action trajectories and synthesize corresponding observations under novel egocentric viewpoints. To address this challenge, we present EgoDemoGen, a framework that generates paired observation--action demonstrations under novel egocentric viewpoints through two key components: 1{)} EgoTrajTransfer, which transfers robot trajectories to the novel egocentric coordinate frame through motion-skill segmentation, geometry-aware transformation, and inverse kinematics filtering; and 2{)} EgoViewTransfer, a conditional video generation model that fuses a novel-viewpoint reprojected scene video and a robot motion video rendered from the transferred trajectory to synthesize photorealistic observations, trained with a self-supervised double reprojection strategy without requiring multi-viewpoint data. Experiments in simulation and real-world settings show that EgoDemoGen consistently improves policy success rates under both standard and novel egocentric viewpoints, with absolute gains of +24.6\% and +16.9\% in simulation and +16.0\% and +23.0\% on the real robot. Moreover, EgoViewTransfer achieves superior video generation quality for novel egocentric observations.

cs.RO↗
arXiv2026

Autonomous Agent-Orchestrated Digital Twins (AADT): Leveraging the OpenClaw Framework for State Synchronization in Rare Genetic Disorders

Background: Medical Digital Twins (MDTs) are computational representations of individual patients that integrate clinical, genomic, and physiological data to support diagnosis, treatment planning, and outcome prediction. However, most MDTs remain static or passively updated, creating a critical synchronization gap, especially in rare genetic disorders where phenotypes, genomic interpretations, and care guidelines evolve over time. Methods: We propose an agent-orchestrated digital twin framework using OpenClaw's proactive "heartbeat" mechanism and modular Agent Skills. This Autonomous Agent-orchestrated Digital Twin (AADT) system continuously monitors local and external data streams (e.g., patient-reported phenotypes and updates in variant classification databases) and executes automated workflows for data ingestion, normalization, state updates, and trigger-based analysis. Results: A prototype implementation demonstrates that agent orchestration can continuously synchronize MDT states with both longitudinal phenotype updates and evolving genomic knowledge. In rare disease settings, this enables earlier diagnosis and more accurate modeling of disease progression. We present two case studies, including variant reinterpretation and longitudinal phenotype tracking, highlighting how AADTs support timely, auditable updates for both research and clinical care. Conclusion: The AADT framework addresses the key bottleneck of real-time synchronization in MDTs, enabling scalable and continuously updated patient models. We also discuss data security considerations and mitigation strategies through human-in-the-loop system design.

q-bio.QM↗