arXiv subjects

Kun Zhao

Publications and source records attributed to Kun Zhao.

At least 37 records · Page 2Linked to original sources

arXiv2026

Listen, Look, Drive: Coupling Audio Instructions for User-aware VLA-based Autonomous Driving

Vision Language Action (VLA) models promise an open-vocabulary interface that can translate perceptual ambiguity into semantically grounded driving decisions, yet they still treat language as a static prior fixed at inference time. As a result, the model must infer continuously shifting objectives from pixels alone, yielding delayed or overly conservative maneuvers. We argue that effective VLAs for autonomous driving need an online channel in which users can influence driving with specific intentions. To this end, we present EchoVLA, a user-aware VLA that couples camera streams with in situ audio instructions. We augment the nuScenes dataset with temporally aligned, intent-specific speech commands generated by converting ego-motion descriptions into synthetic audios. Further, we compose emotional speech-trajectory pairs into a multimodal Chain-of-Thought (CoT) for fine-tuning a Multimodal Large Model (MLM) based on Qwen2.5-Omni. Specifically, we synthesize the audio-augmented dataset with different emotion types paired with corresponding driving behaviors, leveraging the emotional cues embedded in tone, pitch, and speech tempo to reflect varying user states, such as urgent or hesitant intentions, thus enabling our EchoVLA to interpret not only the semantic content but also the emotional context of audio commands for more nuanced and emotionally adaptive driving behavior. In open-loop benchmarks, our approach reduces the average L2 error by $59.4\%$ and the collision rate by $74.4\%$ compared to the baseline of vision-only perception. More experiments on nuScenes dataset validate that EchoVLA not only steers the trajectory through audio instructions, but also modulates driving behavior in response to the emotions detected in the user's speech.

eess.AS↗

arXiv2026

GRAPHMOE: Amplifying Cognitive Depth of Mixture-of-Experts Network via Introducing Self-Rethinking Mechanism

Traditional Mixture-of-Experts (MoE) networks benefit from utilizing multiple smaller expert models as opposed to a single large network. However, these experts typically operate independently, leaving a question open about whether interconnecting these models could enhance the performance of MoE networks. In response, we introduce GRAPHMOE, a novel method aimed at augmenting the cognitive depth of language models via a self-rethinking mechanism constructed on Pseudo GraphMoE networks. GRAPHMOE employs a recurrent routing strategy to simulate iterative thinking steps, thereby facilitating the flow of information among expert nodes. We implement the GRAPHMOE architecture using Low-Rank Adaptation techniques (LoRA) and conduct extensive experiments on various benchmark datasets. The experimental results reveal that GRAPHMOE outperforms other LoRA based models, achieving state-of-the-art (SOTA) performance. Additionally, this study explores a novel recurrent routing strategy that may inspire further advancements in enhancing the reasoning capabilities of language models.

cs.CL↗

arXiv2025

Unveiling Explicit Patterns: Exact Steady States and Stability in a Confined Chemotaxis Model

Inspired by Carrillo-Li-Wang's work [Proc. London Math. Soc., 2021] on stationary solutions to the singular Keller-Segel system, this paper presents a novel family of explicit steady-state solutions for the same model on a bounded interval, expressed in terms of trigonometric and hyperbolic functions. Under Dirichlet boundary conditions and within a biologically stable parameter regime, these solutions, including singular types such as secant and cosecant, are rigorously derived and analyzed. Their stability is established via energy methods, yielding precise thresholds for pattern persistence. These results provide valuable benchmarks for numerical validation and offer insights into boundary-driven pattern formation.

math.AP↗

arXiv2025

R-GenIMA: Integrating Neuroimaging and Genetics with Interpretable Multimodal AI for Alzheimer's Disease Progression

Early detection of Alzheimer's disease (AD) requires models capable of integrating macro-scale neuroanatomical alterations with micro-scale genetic susceptibility, yet existing multimodal approaches struggle to align these heterogeneous signals. We introduce R-GenIMA, an interpretable multimodal large language model that couples a novel ROI-wise vision transformer with genetic prompting to jointly model structural MRI and single nucleotide polymorphisms (SNPs) variations. By representing each anatomically parcellated brain region as a visual token and encoding SNP profiles as structured text, the framework enables cross-modal attention that links regional atrophy patterns to underlying genetic factors. Applied to the ADNI cohort, R-GenIMA achieves state-of-the-art performance in four-way classification across normal cognition (NC), subjective memory concerns (SMC), mild cognitive impairment (MCI), and AD. Beyond predictive accuracy, the model yields biologically meaningful explanations by identifying stage-specific brain regions and gene signatures, as well as coherent ROI-Gene association patterns across the disease continuum. Attention-based attribution revealed genes consistently enriched for established GWAS-supported AD risk loci, including APOE, BIN1, CLU, and RBFOX1. Stage-resolved neuroanatomical signatures identified shared vulnerability hubs across disease stages alongside stage-specific patterns: striatal involvement in subjective decline, frontotemporal engagement during prodromal impairment, and consolidated multimodal network disruption in AD. These results demonstrate that interpretable multimodal AI can synthesize imaging and genetics to reveal mechanistic insights, providing a foundation for clinically deployable tools that enable earlier risk stratification and inform precision therapeutic strategies in Alzheimer's disease.

cs.LG↗

arXiv2025

Why Text Prevails: Vision May Undermine Multimodal Medical Decision Making

With the rapid progress of large language models (LLMs), advanced multimodal large language models (MLLMs) have demonstrated impressive zero-shot capabilities on vision-language tasks. In the biomedical domain, however, even state-of-the-art MLLMs struggle with basic Medical Decision Making (MDM) tasks. We investigate this limitation using two challenging datasets: (1) three-stage Alzheimer's disease (AD) classification (normal, mild cognitive impairment, dementia), where category differences are visually subtle, and (2) MIMIC-CXR chest radiograph classification with 14 non-mutually exclusive conditions. Our empirical study shows that text-only reasoning consistently outperforms vision-only or vision-text settings, with multimodal inputs often performing worse than text alone. To mitigate this, we explore three strategies: (1) in-context learning with reason-annotated exemplars, (2) vision captioning followed by text-only inference, and (3) few-shot fine-tuning of the vision tower with classification supervision. These findings reveal that current MLLMs lack grounded visual understanding and point to promising directions for improving multimodal decision making in healthcare.

cs.CV↗

arXiv2025

Transformed $\ell_1$ Regularizations for Robust Principal Component Analysis: Toward a Fine-Grained Understanding

Robust Principal Component Analysis (RPCA) aims to recover a low-rank structure from noisy, partially observed data that is also corrupted by sparse, potentially large-magnitude outliers. Traditional RPCA models rely on convex relaxations, such as nuclear norm and $\ell_1$ norm, to approximate the rank of a matrix and the $\ell_0$ functional (the number of non-zero elements) of another. In this work, we advocate a nonconvex regularization method, referred to as transformed $\ell_1$ (TL1), to improve both approximations. The rationale is that by varying the internal parameter of TL1, its behavior asymptotically approaches either $\ell_0$ or $\ell_1$. Since the rank is equal to the number of non-zero singular values and the nuclear norm is defined as their sum, applying TL1 to the singular values can approximate either the rank or the nuclear norm, depending on its internal parameter. We conduct a fine-grained theoretical analysis of statistical convergence rates, measured in the Frobenius norm, for both the low-rank and sparse components under general sampling schemes. These rates are comparable to those of the classical RPCA model based on the nuclear norm and $\ell_1$ norm. Moreover, we establish constant-order upper bounds on the estimated rank of the low-rank component and the cardinality of the sparse component in the regime where TL1 behaves like $\ell_0$, assuming that the respective matrices are exactly low-rank and exactly sparse. Extensive numerical experiments on synthetic data and real-world applications demonstrate that the proposed approach achieves higher accuracy than the classic convex model, especially under non-uniform sampling schemes.

stat.ML↗

arXiv2025

Iterative projection method for unsteady Navier-Stokes equations with high Reynolds numbers

A new iterative projection method is proposed to solve the unsteady Navier-Stokes equations with high Reynolds numbers. The convectional projection method attempts to project the intermediate velocity to the divergence free space only once per time step. However, such a velocity is not genuinely divergence free in general practice, which can yield large errors when the Reynolds number is high. The new method has several important features: the BDF2 time discretization, the skew-symmetric convection in a semi-implicit form, two modulating parameters, and the iterative projections in each time step. A major difficulty in the proof of iteration convergence is the nonlinear convection. We solve this problem by first analyzing the non-convective scheme with a focus on the spectral properties of the iterative matrix, and then employing a delicate perturbation analysis for the convective scheme. The work achieves the weakly divergence free velocity (strongly divergence free for divergence free finite element spaces), and the rigorous stability and error analysis when the iterations converge. The three dimensional numerical tests confirm that this new method can effectively treat high Reynolds numbers with only a few iterations per time step, where the convectional projection method and the iterative projection method with the explicit convection would fail.

math.NA↗

arXiv2025

Emphasising Structured Information: Integrating Abstract Meaning Representation into LLMs for Enhanced Open-Domain Dialogue Evaluation

Automatic open-domain dialogue evaluation has attracted increasing attention, yet remains challenging due to the complexity of assessing response appropriateness. Traditional evaluation metrics, typically trained with true positive and randomly selected negative responses, tend to assign higher scores to responses that share greater content similarity with contexts. However, adversarial negative responses, despite possessing high lexical overlap with contexts, can be semantically incongruous. Consequently, existing metrics struggle to effectively evaluate such responses, resulting in low correlations with human judgments. While recent studies have demonstrated the effectiveness of Large Language Models (LLMs) for open-domain dialogue evaluation, they still face challenges in handling adversarial negative examples. We propose a novel evaluation framework that integrates Abstract Meaning Representation (AMR) enhanced domain-specific language models (SLMs) with LLMs. Our SLMs explicitly incorporate AMR graph information through a gating mechanism for enhanced semantic representation learning, while both SLM predictions and AMR knowledge are integrated into LLM prompts for robust evaluation. Extensive experiments on open-domain dialogue evaluation tasks demonstrate the superiority of our method compared to state-of-the-art baselines. Our comprehensive ablation studies reveal that AMR graph information contributes substantially more to performance improvements. Our framework achieves strong correlations with human judgments across multiple datasets, establishing a new benchmark for dialogue evaluation. Our code and data are publicly available.

cs.CL↗

arXiv2025

Crafting Customisable Characters with LLMs: A Persona-Driven Role-Playing Agent Framework

Large Language Models (LLMs) demonstrate remarkable ability to comprehend instructions and generate human-like text, enabling sophisticated agent simulation beyond basic behavior replication. However, the potential for creating freely customisable characters remains underexplored. We introduce the Customisable Conversation Agent Framework, which employs LLMs to simulate real-world characters through personalised characteristic feature injection, enabling diverse character creation according to user preferences. We propose the SimsConv dataset, comprising 68 customised characters and 13,971 multi-turn role-playing dialogues across 1,360 real-world scenes. Characters are initially customised using pre-defined elements (career, aspiration, traits, skills), then expanded through personal and social profiles. Building on this, we present SimsChat, a freely customisable role-playing agent incorporating various realistic settings and topic-specified character interactions. Experimental results on both SimsConv and WikiRoleEval datasets demonstrate SimsChat's superior performance in maintaining character consistency, knowledge accuracy, and appropriate question rejection compared to existing models. Our framework provides valuable insights for developing more accurate and customisable human simulacra. Our data and code are publicly available at https://github.com/Bernard-Yang/SimsChat.

cs.CL↗

arXiv2025

DRE: An Effective Dual-Refined Method for Integrating Small and Large Language Models in Open-Domain Dialogue Evaluation

Large Language Models (LLMs) excel at many tasks but struggle with ambiguous scenarios where multiple valid responses exist, often yielding unreliable results. Conversely, Small Language Models (SLMs) demonstrate robustness in such scenarios but are susceptible to misleading or adversarial inputs. We observed that LLMs handle negative examples effectively, while SLMs excel with positive examples. To leverage their complementary strengths, we introduce SLIDE (Small and Large Integrated for Dialogue Evaluation), a method integrating SLMs and LLMs via adaptive weighting. Building on SLIDE, we further propose a Dual-Refinement Evaluation (DRE) method to enhance SLM-LLM integration: (1) SLM-generated insights guide the LLM to produce initial evaluations; (2) SLM-derived adjustments refine the LLM's scores for improved accuracy. Experiments demonstrate that DRE outperforms existing methods, showing stronger alignment with human judgment across diverse benchmarks. This work illustrates how combining small and large models can yield more reliable evaluation tools, particularly for open-ended tasks such as dialogue evaluation.

cs.CL↗

arXiv2025

X-ray Made Simple: Lay Radiology Report Generation and Robust Evaluation

Radiology Report Generation (RRG) has advanced considerably with the development of multimodal generative models. Despite the progress, the field still faces significant challenges in evaluation, as existing metrics lack robustness and fairness. We reveal that, RRG with high performance on existing lexical-based metrics (e.g. BLEU) might be more of a mirage - a model can get a high BLEU only by learning the template of reports. This has become a pressing issue for RRG due to the highly patternized nature of these reports. In addition, standard radiology reports are often highly technical. Helping patients understand these reports is crucial from a patient's perspective, yet this has been largely overlooked in previous work. In this work, we un-intuitively approach these problems by proposing the Layman's RRG framework that can systematically improve RRG with day-to-day language. Specifically, our framework first contributes a translated Layman's terms dataset. Building upon the dataset, we then propose a semantics-based evaluation method, which is effective in mitigating the inflated numbers of BLEU and provides more robust evaluation. We show that training on the layman's terms dataset encourages models to focus on the semantics of the reports, as opposed to overfitting to learning the report templates. Last, we reveal a promising scaling law between the number of training examples and semantics gain provided by our dataset, compared to the inverse pattern brought by the original formats.

cs.CL↗

arXiv2025

Atomistic mechanisms of dynamics in a two-dimensional dodecagonal quasicrystal

Quasicrystals have been observed in a variety of materials ranging from metal alloys to block copolymers. However, their structural and dynamical properties cannot be readily described in terms of conventional solid-state models of liquids and solids. We may expect the dynamics of this specific class of quasicrystalline materials to be more like glass-forming liquids in the sense of exhibiting large fluctuations in the local mobility ("dynamic heterogeneity") and non-Arrhenius temperature dependence of relaxation and diffusion. In this work, we investigate a model dodecagonal quasicrystal material in two dimensions (2D) using molecular dynamics (MD) simulations, with a focus on heterogeneous dynamics and non-Arrhenius relaxation and diffusion. As observed in glass-forming liquids and heated crystals, we observe a two-stage relaxation dynamics in the self-intermediate scattering function $F_s(k,t)$ of our quasicrystal material. It involves a fast $β$-relaxation and $α$ relaxation process having a highly temperature dependent relaxation time whose activation energy varies in concert with the extent of string-like collective motion, a phenomenon recognized to occur in glass-forming liquids at low temperatures and crystalline materials at elevated temperatures. After examining the dynamics of our dodecagonal quasicrystalline material in great detail, we conclude that the dynamics of these materials more closely resembles observations on metallic glass-forming liquids than crystalline materials.

cond-mat.mtrl-sci↗

arXiv2025

LMFormer: Lane based Motion Prediction Transformer

Motion prediction plays an important role in autonomous driving. This study presents LMFormer, a lane-aware transformer network for trajectory prediction tasks. In contrast to previous studies, our work provides a simple mechanism to dynamically prioritize the lanes and shows that such a mechanism introduces explainability into the learning behavior of the network. Additionally, LMFormer uses the lane connection information at intersections, lane merges, and lane splits, in order to learn long-range dependency in lane structure. Moreover, we also address the issue of refining the predicted trajectories and propose an efficient method for iterative refinement through stacked transformer layers. For benchmarking, we evaluate LMFormer on the nuScenes dataset and demonstrate that it achieves SOTA performance across multiple metrics. Furthermore, the Deep Scenario dataset is used to not only illustrate cross-dataset network performance but also the unification capabilities of LMFormer to train on multiple datasets and achieve better performance.

cs.CV↗

arXiv2025

FantasyTalking: Realistic Talking Portrait Generation via Coherent Motion Synthesis

Creating a realistic animatable avatar from a single static portrait remains challenging. Existing approaches often struggle to capture subtle facial expressions, the associated global body movements, and the dynamic background. To address these limitations, we propose a novel framework that leverages a pretrained video diffusion transformer model to generate high-fidelity, coherent talking portraits with controllable motion dynamics. At the core of our work is a dual-stage audio-visual alignment strategy. In the first stage, we employ a clip-level training scheme to establish coherent global motion by aligning audio-driven dynamics across the entire scene, including the reference portrait, contextual objects, and background. In the second stage, we refine lip movements at the frame level using a lip-tracing mask, ensuring precise synchronization with audio signals. To preserve identity without compromising motion flexibility, we replace the commonly used reference network with a facial-focused cross-attention module that effectively maintains facial consistency throughout the video. Furthermore, we integrate a motion intensity modulation module that explicitly controls expression and body motion intensity, enabling controllable manipulation of portrait movements beyond mere lip motion. Extensive experimental results show that our proposed approach achieves higher quality with better realism, coherence, motion intensity, and identity preservation. Ours project page: https://fantasy-amap.github.io/fantasy-talking/.

cs.CV↗

arXiv2025

One Snapshot is All You Need: A Generalized Method for mmWave Signal Generation

Wireless sensing systems, particularly those using mmWave technology, offer distinct advantages over traditional vision-based approaches, such as enhanced privacy and effectiveness in poor lighting conditions. These systems, leveraging FMCW signals, have shown success in human-centric applications like localization, gesture recognition, and so on. However, comprehensive mmWave datasets for diverse applications are scarce, often constrained by pre-processed signatures (e.g., point clouds or RA heatmaps) and inconsistent annotation formats. To overcome these limitations, we propose mmGen, a novel and generalized framework tailored for full-scene mmWave signal generation. By constructing physical signal transmission models, mmGen synthesizes human-reflected and environment-reflected mmWave signals from the constructed 3D meshes. Additionally, we incorporate methods to account for material properties, antenna gains, and multipath reflections, enhancing the realism of the synthesized signals. We conduct extensive experiments using a prototype system with commercial mmWave devices and Kinect sensors. The results show that the average similarity of Range-Angle and micro-Doppler signatures between the synthesized and real-captured signals across three different environments exceeds 0.91 and 0.89, respectively, demonstrating the effectiveness and practical applicability of mmGen.

cs.CV↗

arXiv2025

Noisy Low-Rank Matrix Completion via Transformed $L_1$ Regularization and its Theoretical Properties

This paper focuses on recovering an underlying matrix from its noisy partial entries, a problem commonly known as matrix completion. We delve into the investigation of a non-convex regularization, referred to as transformed $L_1$ (TL1), which interpolates between the rank and the nuclear norm of matrices through a hyper-parameter $a \in (0, \infty)$. While some literature adopts such regularization for matrix completion, it primarily addresses scenarios with uniformly missing entries and focuses on algorithmic advances. To fill in the gap in the current literature, we provide a comprehensive statistical analysis for the estimator from a TL1-regularized recovery model under general sampling distribution. In particular, we show that when $a$ is sufficiently large, the matrix recovered by the TL1-based model enjoys a convergence rate measured by the Frobenius norm, comparable to that of the model based on the nuclear norm, despite the challenges posed by the non-convexity of the TL1 regularization. When $a$ is small enough, we show that the rank of the estimated matrix remains a constant order when the true matrix is exactly low-rank. A trade-off between controlling the error and the rank is established through different choices of tuning parameters. The appealing practical performance of TL1 regularization is demonstrated through a simulation study that encompasses various sampling mechanisms, as well as two real-world applications. Additionally, the role of the hyper-parameter $a$ on the TL1-based model is explored via experiments to offer guidance in practical scenarios.

math.ST↗

arXiv2025

Stationary solutions with vacuum for a hyperbolic-parabolic chemotaxis model in dimension two

In this research, we study the existence of stationary solutions with vacuum to a hyperbolic-parabolic chemotaxis model with nonlinear pressure in dimension two that describes vasculogenesis. We seek solutions in the radial symmetric class of the whole space, in which the system will be reduced to a system of ODE's on $(0,\infty)$. The fundamental solutions to the ODE system are the Bessel functions of different types. We find two nontrivial solutions. One is formed by half bump (positive density region) starting at $r=0$ and a region of vacuum on the right. Another one is a full nonsymmetric bump away from $r=0$. These solutions bear certain resemblance to in vitro vascular network and the numerically produced structure by Gamba et al arXiv:cond-mat/0303468v1. We also show the nonexistence of full bump starting at $r=0$ and nonexistence of full symmetric bump away from $r=0$.

math.AP↗

arXiv2025

Exploring EEG and Eye Movement Fusion for Multi-Class Target RSVP-BCI

Rapid Serial Visual Presentation (RSVP)-based Brain-Computer Interfaces (BCIs) facilitate high-throughput target image detection by identifying event-related potentials (ERPs) evoked in EEG signals. The RSVP-BCI systems effectively detect single-class targets within a stream of images but have limited applicability in scenarios that require detecting multiple target categories. Multi-class RSVP-BCI systems address this limitation by simultaneously identifying the presence of a target and distinguishing its category. However, existing multi-class RSVP decoding algorithms predominantly rely on single-modality EEG decoding, which restricts their performance improvement due to the high similarity between ERPs evoked by different target categories. In this work, we introduce eye movement (EM) modality into multi-class RSVP decoding and explore EEG and EM fusion to enhance decoding performance. First, we design three independent multi-class target RSVP tasks and build an open-source dataset comprising EEG and EM signals from 43 subjects. Then, we propose the Multi-class Target RSVP EEG and EM fusion Network (MTREE-Net) to enhance multi-class RSVP decoding. Specifically, a dual-complementary module is proposed to strengthen the differentiation of uni-modal features across categories. To improve multi-modal fusion performance, we adopt a dynamic reweighting fusion strategy guided by theoretically derived modality contribution ratios. Furthermore, we reduce the misclassification of non-target samples through knowledge transfer between two hierarchical classifiers. Extensive experiments demonstrate the feasibility of integrating EM signals into multi-class RSVP decoding and highlight the superior performance of MTREE-Net compared to existing RSVP decoding methods. The proposed MTREE-Net and open-source dataset provide a promising framework for developing practical multi-class RSVP-BCI systems.

cs.HC↗