Yuan Zhao

Yue Wang, Yuan Zhao, Shengcheng Yu et al.

Agentic AI is increasingly being integrated into software engineering workflows. In crowdsourced testing, however, the large volume and uneven quality of submitted reports still create a substantial review burden for developers. In prior work, we developed and validated a multi-agent assessment backbone based on the LLM-as-a-Judge paradigm. That backbone assesses reports along three dimensions--textuality, adequacy, and competitiveness--and was shown to align well with human consensus while substantially reducing assessment effort. Yet reliable automated judging does not by itself show whether agent outputs can improve human work when embedded into workflow. This paper studies that missing question in the context of crowdsourced testing. We investigate whether assessment-derived, actionable feedback can improve how testers revise reports, perform on later tasks, and transfer reporting practices across applications. To do so, we conducted a controlled four-stage human-subject study with 20 testers across three real-world applications. The results show that agent-generated feedback supports immediate improvements in revised reports, better first submissions on a new task after prior feedback exposure, and evidence of partial but meaningful transfer to a later application. A post-task questionnaire completed by 17 participants complements these artifact-based findings by suggesting that the feedback was generally understandable, acted upon in revision, and carried into later tasks, while also revealing remaining friction in specificity and execution. Overall, the study provides empirical evidence that, in the studied crowdsourced testing setting, assessment agents can serve not only as post-hoc judges but also as workflow-integrated feedback providers that support upstream report-quality improvement.

Yuan Zhao, Zhenqi Jia, Rui Liu et al.

Automatic Video Dubbing (AVD) aims to take the given script and generate speech that aligns with lip motion and prosody expressiveness. Current AVD models mainly utilize visual information of the current sentence to enhance the prosody of synthesized speech. However, it is crucial to consider whether the prosody of the generated dubbing aligns with the multimodal context, as the dubbing will be combined with the original context in the final video. This aspect has been overlooked in previous studies. To address this issue, we propose a Multimodal Context-aware video Dubbing model, termed \textbf{MCDubber}, to convert the modeling object from a single sentence to a longer sequence with context information to ensure the consistency of the global context prosody. MCDubber comprises three main components: (1) A context duration aligner aims to learn the context-aware alignment between the text and lip frames; (2) A context prosody predictor seeks to read the global context visual sequence and predict the context-aware global energy and pitch; (3) A context acoustic decoder ultimately predicts the global context mel-spectrogram with the assistance of adjacent ground-truth mel-spectrograms of the target sentence. Through this process, MCDubber fully considers the influence of multimodal context on the prosody expressiveness of the current sentence when dubbing. The extracted mel-spectrogram belonging to the target sentence from the output context mel-spectrograms is the final required dubbing audio. Extensive experiments on the Chem benchmark dataset demonstrate that our MCDubber significantly improves dubbing expressiveness compared to all advanced baselines. The code and demos are available at https://github.com/XiaoYuanJun-zy/MCDubber.

Huajie Chen, Tianqing Zhu, Yuan Zhao et al.

Image deep steganography (IDS) is a technique that utilizes deep learning to embed a secret image invisibly into a cover image to generate a container image. However, the container images generated by convolutional neural networks (CNNs) are vulnerable to attacks that distort their high-frequency components. To address this problem, we propose a novel method called Low-frequency Image Deep Steganography (LIDS) that allows frequency distribution manipulation in the embedding process. LIDS extracts a feature map from the secret image and adds it to the cover image to yield the container image. The container image is not directly output by the CNNs, and thus, it does not contain high-frequency artifacts. The extracted feature map is regulated by a frequency loss to ensure that its frequency distribution mainly concentrates on the low-frequency domain. To further enhance robustness, an attack layer is inserted to damage the container image. The retrieval network then retrieves a recovered secret image from a damaged container image. Our experiments demonstrate that LIDS outperforms state-of-the-art methods in terms of robustness, while maintaining high fidelity and specificity. By avoiding high-frequency artifacts and manipulating the frequency distribution of the embedded feature map, LIDS achieves improved robustness against attacks that distort the high-frequency components of container images.

Matthew Dowling, Yuan Zhao, Il Memming Park

Latent Gaussian process (GP) models are widely used in neuroscience to uncover hidden state evolutions from sequential observations, mainly in neural activity recordings. While latent GP models provide a principled and powerful solution in theory, the intractable posterior in non-conjugate settings necessitates approximate inference schemes, which may lack scalability. In this work, we propose cvHM, a general inference framework for latent GP models leveraging Hida-Matérn kernels and conjugate computation variational inference (CVI). With cvHM, we are able to perform variational inference of latent neural trajectories with linear time complexity for arbitrary likelihoods. The reparameterization of stationary kernels using Hida-Matérn GPs helps us connect the latent variable models that encode prior assumptions through dynamical systems to those that encode trajectory assumptions through GPs. In contrast to previous work, we use bidirectional information filtering, leading to a more concise implementation. Furthermore, we employ the Whittle approximate likelihood to achieve highly efficient hyperparameter learning.

Tingyu Jiang, Shen Li, Yiyao Song et al.

Instruction tuning plays a critical role in enhancing the performance and efficiency of Large Language Models (LLMs). Its success depends not only on the quality of the instruction data but also on the inherent capabilities of the LLM itself. Some studies suggest that even a small amount of high-quality data can achieve instruction fine-tuning results that are on par with, or even exceed, those from using a full-scale dataset. However, rather than focusing solely on calculating data quality scores to evaluate instruction data, there is a growing need to select high-quality data that maximally enhances the performance of instruction tuning for a given LLM. In this paper, we propose the Model Instruction Weakness Value (MIWV) as a novel metric to quantify the importance of instruction data in enhancing model's capabilities. The MIWV metric is derived from the discrepancies in the model's responses when using In-Context Learning (ICL), helping identify the most beneficial data for enhancing instruction tuning performance. Our experimental results demonstrate that selecting only the top 1\% of data based on MIWV can outperform training on the full dataset. Furthermore, this approach extends beyond existing research that focuses on data quality scoring for data selection, offering strong empirical evidence supporting the effectiveness of our proposed method.

Yuan Zhao, Youwei Pang, Jiaming Zuo et al.

Recent progress in promptable segmentation has shifted visual perception from object-level localization toward concept-level understanding. However, the notion of a concept remains under-specified, making it unclear whether current methods truly generalize beyond category recognition. In this work, we formalize generalized concept segmentation through a three-level taxonomy consisting of context-independent (CI), context-dependent (CD), and context-reasoning (CR) concepts, which reveals a clear capability gap across increasing levels of cognitive complexity. To address this challenge, we propose ConceptSeg-R1, a unified framework that reformulates concept segmentation as rule-induced concept grounding. At the core of our method is Meta-GRPO, a meta-reinforcement learning mechanism that learns transferable task rules from visual demonstrations and verifies them through proxy reasoning. The inferred reasoning states are then translated into segmentation-ready concept prompts via a lightweight concept translation module, enabling deductive application to target images. A shortcut routing strategy further preserves the native efficiency of segmentation models on simple cases. To systematically evaluate generalized concept segmentation, we conduct extensive experiments across diverse CI, CD, and CR concept segmentation benchmarks spanning natural, industrial, medical and reasoning-intensive domains. Without bells and whistles, ConceptSeg-R1 achieves strong performance across the full concept hierarchy while maintaining the native capability of promptable segmentation backbones. As an initial step toward segmenting any concept, we hope ConceptSeg-R1 can serve as a practical baseline for advancing segmentation from object-level prediction toward concept-level understanding.

Kai Yao, Penglei Gao, Lichun Li et al.

Parameter-Efficient Fine-Tuning (PEFT) methods have gained significant popularity for adapting pre-trained Large Language Models (LLMs) to downstream tasks, primarily due to their potential to significantly reduce memory and computational overheads. However, a common limitation in most PEFT approaches is their application of a uniform architectural design across all layers. This uniformity involves identical trainable modules and ignores the varying importance of each layer, leading to sub-optimal fine-tuning results. To overcome the above limitation and obtain better performance, we develop a novel approach, Importance-aware Sparse Tuning (IST), to fully utilize the inherent sparsity and select the most important subset of full layers with effective layer-wise importance scoring. The proposed IST is a versatile and plug-and-play technique compatible with various PEFT methods that operate on a per-layer basis. By leveraging the estimated importance scores, IST dynamically updates these selected layers in PEFT modules, leading to reduced memory demands. We further provide theoretical proof of convergence and empirical evidence of superior performance to demonstrate the advantages of IST over uniform updating strategies. Extensive experiments on a range of LLMs, PEFTs, and downstream tasks substantiate the effectiveness of our proposed method, showcasing IST's capacity to enhance existing layer-based PEFT methods. Our code is available at https://github.com/Kaiseem/IST.

Yuan Zhao, Rui Liu, Gaoxiang Cong

Automatic Video Dubbing (AVD) generates speech aligned with lip motion and facial emotion from scripts. Recent research focuses on modeling multimodal context to enhance prosody expressiveness but overlooks two key issues: 1) Multiscale prosody expression attributes in the context influence the current sentence's prosody. 2) Prosody cues in context interact with the current sentence, impacting the final prosody expressiveness. To tackle these challenges, we propose M2CI-Dubber, a Multiscale Multimodal Context Interaction scheme for AVD. This scheme includes two shared M2CI encoders to model the multiscale multimodal context and facilitate its deep interaction with the current sentence. By extracting global and local features for each modality in the context, utilizing attention-based mechanisms for aggregation and interaction, and employing an interaction-based graph attention network for fusion, the proposed approach enhances the prosody expressiveness of synthesized speech for the current sentence. Experiments on the Chem dataset show our model outperforms baselines in dubbing expressiveness. The code and demos are available at \textcolor[rgb]{0.93,0.0,0.47}{https://github.com/AI-S2-Lab/M2CI-Dubber}.

Zhuo Zhang, Xiong Xiong, Sen Zhang et al.

PDEs arise ubiquitously in science and engineering, where solutions depend on parameters (physical properties, boundary conditions, geometry). Traditional numerical methods require re-solving the PDE for each parameter, making parameter space exploration prohibitively expensive. Recent machine learning advances, particularly physics-informed neural networks (PINNs) and neural operators, have revolutionized parametric PDE solving by learning solution operators that generalize across parameter spaces. We critically analyze two main paradigms: (1) PINNs, which embed physical laws as soft constraints and excel at inverse problems with sparse data, and (2) neural operators (e.g., DeepONet, Fourier Neural Operator), which learn mappings between infinite-dimensional function spaces and achieve unprecedented generalization. Through comparisons across fluid dynamics, solid mechanics, heat transfer, and electromagnetics, we show neural operators can achieve computational speedups of $10^3$ to $10^5$ times faster than traditional solvers for multi-query scenarios, while maintaining comparable accuracy. We provide practical guidance for method selection, discuss theoretical foundations (universal approximation, convergence), and identify critical open challenges: high-dimensional parameters, complex geometries, and out-of-distribution generalization. This work establishes a unified framework for understanding parametric PDE solvers via operator learning, offering a comprehensive, incrementally updated resource for this rapidly evolving field

Rui Liu, Yuan Zhao, Zhenqi Jia

The automatic movie dubbing model generates vivid speech from given scripts, replicating a speaker's timbre from a brief timbre prompt while ensuring lip-sync with the silent video. Existing approaches simulate a simplified workflow where actors dub directly without preparation, overlooking the critical director-actor interaction. In contrast, authentic workflows involve a dynamic collaboration: directors actively engage with actors, guiding them to internalize the context cues, specifically emotion, before performance. To address this issue, we propose a new Retrieve-Augmented Director-Actor Interaction Learning scheme to achieve authentic movie dubbing, termed Authentic-Dubber, which contains three novel mechanisms: (1) We construct a multimodal Reference Footage library to simulate the learning footage provided by directors. Note that we integrate Large Language Models (LLMs) to achieve deep comprehension of emotional representations across multimodal signals. (2) To emulate how actors efficiently and comprehensively internalize director-provided footage during dubbing, we propose an Emotion-Similarity-based Retrieval-Augmentation strategy. This strategy retrieves the most relevant multimodal information that aligns with the target silent video. (3) We develop a Progressive Graph-based speech generation approach that incrementally incorporates the retrieved multimodal emotional knowledge, thereby simulating the actor's final dubbing process. The above mechanisms enable the Authentic-Dubber to faithfully replicate the authentic dubbing workflow, achieving comprehensive improvements in emotional expressiveness. Both subjective and objective evaluations on the V2C Animation benchmark dataset validate the effectiveness. The code and demos are available at https://github.com/AI-S2-Lab/Authentic-Dubber.

Yuan Zhao, Youwei Pang, Lihe Zhang et al.

Existing anomaly detection (AD) methods often treat the modality and class as independent factors. Although this paradigm has enriched the development of AD research branches and produced many specialized models, it has also led to fragmented solutions and excessive memory overhead. Moreover, reconstruction-based multi-class approaches typically rely on shared decoding paths, which struggle to handle large variations across domains, resulting in distorted normality boundaries, domain interference, and high false alarm rates. To address these limitations, we propose UniMMAD, a unified framework for multi-modal and multi-class anomaly detection. At the core of UniMMAD is a Mixture-of-Experts (MoE)-driven feature decompression mechanism, which enables adaptive and disentangled reconstruction tailored to specific domains. This process is guided by a ``general to specific'' paradigm. In the encoding stage, multi-modal inputs of varying combinations are compressed into compact, general-purpose features. The encoder incorporates a feature compression module to suppress latent anomalies, encourage cross-modal interaction, and avoid shortcut learning. In the decoding stage, the general features are decompressed into modality-specific and class-specific forms via a sparsely-gated cross MoE, which dynamically selects expert pathways based on input modality and class. To further improve efficiency, we design a grouped dynamic filtering mechanism and a MoE-in-MoE structure, reducing parameter usage by 75\% while maintaining sparse activation and fast inference. UniMMAD achieves state-of-the-art performance on 9 anomaly detection datasets, spanning 3 fields, 12 modalities, and 66 classes. The source code will be available at https://github.com/yuanzhao-CVLAB/UniMMAD.

Zhuo Zhang, Xi Yang, Ying Miao et al.

While Transformers have demonstrated remarkable potential in modeling Partial Differential Equations (PDEs), modeling large-scale unstructured meshes with complex geometries remains a significant challenge. Existing efficient architectures often employ feature dimensionality reduction strategies, which inadvertently induces Geometric Aliasing, resulting in the loss of critical physical boundary information. To address this, we propose the Physics-Geometry Operator Transformer (PGOT), designed to reconstruct physical feature learning through explicit geometry awareness. Specifically, we propose Spectrum-Preserving Geometric Attention (SpecGeo-Attention). Utilizing a ``physics slicing-geometry injection" mechanism, this module incorporates multi-scale geometric encodings to explicitly preserve multi-scale geometric features while maintaining linear computational complexity $O(N)$. Furthermore, PGOT dynamically routes computations to low-order linear paths for smooth regions and high-order non-linear paths for shock waves and discontinuities based on spatial coordinates, enabling spatially adaptive and high-precision physical field modeling. PGOT achieves consistent state-of-the-art performance across four standard benchmarks and excels in large-scale industrial tasks including airfoil and car designs.

Wenjie Wang, Wei Wu, Ying Liu et al.

Medical document OCR is challenging due to complex layouts, domain-specific terminology, and noisy annotations, while requiring strict field-level exact matching. Existing OCR systems and general-purpose vision-language models often fail to reliably parse such documents. We propose MeDocVL, a post-trained vision-language model for query-driven medical document parsing. Our framework combines Training-driven Label Refinement to construct high-quality supervision from noisy annotations, with a Noise-aware Hybrid Post-training strategy that integrates reinforcement learning and supervised fine-tuning to achieve robust and precise extraction. Experiments on medical invoice benchmarks show that MeDocVL consistently outperforms conventional OCR systems and strong VLM baselines, achieving state-of-the-art performance under noisy supervision.

Matthew Dowling, Yuan Zhao, Il Memming Park

State-space graphical models and the variational autoencoder framework provide a principled apparatus for learning dynamical systems from data. State-of-the-art probabilistic approaches are often able to scale to large problems at the cost of flexibility of the variational posterior or expressivity of the dynamics model. However, those consolidations can be detrimental if the ultimate goal is to learn a generative model capable of explaining the spatiotemporal structure of the data and making accurate forecasts. We introduce a low-rank structured variational autoencoding framework for nonlinear Gaussian state-space graphical models capable of capturing dense covariance structures that are important for learning dynamical systems with predictive capabilities. Our inference algorithm exploits the covariance structures that arise naturally from sample based approximate Gaussian message passing and low-rank amortized posterior updates -- effectively performing approximate variational smoothing with time complexity scaling linearly in the state dimensionality. In comparisons with other deep state-space model architectures our approach consistently demonstrates the ability to learn a more predictive generative model. Furthermore, when applied to neural physiological recordings, our approach is able to learn a dynamical system capable of forecasting population spiking and behavioral correlates from a small portion of single trials.

Xiaoqi Zhao, Youwei Pang, Shijie Chang et al.

As large-scale foundation models trained on billions of image--mask pairs covering a vast diversity of scenes, objects, and contexts, SAM and its upgraded version, SAM~2, have significantly influenced multiple fields within computer vision. Leveraging such unprecedented data diversity, they exhibit strong open-world segmentation capabilities, with SAM~2 further enhancing these capabilities to support high-quality video segmentation. While SAMs (SAM and SAM~2) have demonstrated excellent performance in segmenting context-independent concepts like people, cars, and roads, they overlook more challenging context-dependent (CD) concepts, such as visual saliency, camouflage, industrial defects, and medical lesions. CD concepts rely heavily on global and local contextual information, making them susceptible to shifts in different contexts, which requires strong discriminative capabilities from the model. The lack of comprehensive evaluation of SAMs limits understanding of their performance boundaries, which may hinder the design of future models. In this paper, we conduct a thorough evaluation of SAMs on 11 CD concepts across 2D and 3D images and videos in various visual modalities within natural, medical, and industrial scenes. We develop a unified evaluation framework for SAM and SAM~2 that supports manual, automatic, and intermediate self-prompting, aided by our specific prompt generation and interaction strategies. We further explore the potential of SAM~2 for in-context learning and introduce prompt robustness testing to simulate real-world imperfect prompts. Finally, we analyze the benefits and limitations of SAMs in understanding CD concepts and discuss their future development in segmentation tasks.

Yuan Zhao, Hualei Zhu, Tingyu Jiang et al.

Graphical User Interface (GUI) task automation constitutes a critical frontier in artificial intelligence research. While effective GUI agents synergistically integrate planning and grounding capabilities, current methodologies exhibit two fundamental limitations: (1) insufficient exploitation of cross-model synergies, and (2) over-reliance on synthetic data generation without sufficient utilization. To address these challenges, we propose Co-EPG, a self-iterative training framework for Co-Evolution of Planning and Grounding. Co-EPG establishes an iterative positive feedback loop: through this loop, the planning model explores superior strategies under grounding-based reward guidance via Group Relative Policy Optimization (GRPO), generating diverse data to optimize the grounding model. Concurrently, the optimized Grounding model provides more effective rewards for subsequent GRPO training of the planning model, fostering continuous improvement. Co-EPG thus enables iterative enhancement of agent capabilities through self-play optimization and training data distillation. On the Multimodal-Mind2Web and AndroidControl benchmarks, our framework outperforms existing state-of-the-art methods after just three iterations without requiring external data. The agent consistently improves with each iteration, demonstrating robust self-enhancement capabilities. This work establishes a novel training paradigm for GUI agents, shifting from isolated optimization to an integrated, self-driven co-evolution approach.

Yang Zhang, Chenwei Wang, Ouyang Lu et al.

Vision-Language-Action (VLA) models pre-trained on large, diverse datasets show remarkable potential for general-purpose robotic manipulation. However, a primary bottleneck remains in adapting these models to downstream tasks, especially when the robot's embodiment or the task itself differs from the pre-training data. This discrepancy leads to a significant mismatch in action distributions, demanding extensive data and compute for effective fine-tuning. To address this challenge, we introduce \textbf{Align-Then-stEer (\texttt{ATE})}, a novel, data-efficient, and plug-and-play adaptation framework. \texttt{ATE} first aligns disparate action spaces by constructing a unified latent space, where a variational autoencoder constrained by reverse KL divergence embeds adaptation actions into modes of the pre-training action latent distribution. Subsequently, it steers the diffusion- or flow-based VLA's generation process during fine-tuning via a guidance mechanism that pushes the model's output distribution towards the target domain. We conduct extensive experiments on cross-embodiment and cross-task manipulation in both simulation and real world. Compared to direct fine-tuning of representative VLAs, our method improves the average multi-task success rate by up to \textbf{9.8\%} in simulation and achieves a striking \textbf{32\% success rate gain} in a real-world cross-embodiment setting. Our work presents a general and lightweight solution that greatly enhances the practicality of deploying VLA models to new robotic platforms and tasks.

Kai Yao, Zhaorui Tan, Tiandi Ye et al.

Offsite-tuning is a privacy-preserving method for tuning large language models (LLMs) by sharing a lossy compressed emulator from the LLM owners with data owners for downstream task tuning. This approach protects the privacy of both the model and data owners. However, current offsite tuning methods often suffer from adaptation degradation, high computational costs, and limited protection strength due to uniformly dropping LLM layers or relying on expensive knowledge distillation. To address these issues, we propose ScaleOT, a novel privacy-utility-scalable offsite-tuning framework that effectively balances privacy and utility. ScaleOT introduces a novel layerwise lossy compression algorithm that uses reinforcement learning to obtain the importance of each layer. It employs lightweight networks, termed harmonizers, to replace the raw LLM layers. By combining important original LLM layers and harmonizers in different ratios, ScaleOT generates emulators tailored for optimal performance with various model scales for enhanced privacy protection. Additionally, we present a rank reduction method to further compress the original LLM layers, significantly enhancing privacy with negligible impact on utility. Comprehensive experiments show that ScaleOT can achieve nearly lossless offsite tuning performance compared with full fine-tuning while obtaining better model privacy.

Yuan Zhao, Zhenqi Jia, Yongqiang Zhang

Reference Audio-Visual Segmentation (Ref-AVS) aims to segment objects in audible videos based on multimodal cues in reference expressions. Previous methods overlook the explicit recognition of expression difficulty and dominant modality in multimodal cues, over-rely on the quality of the instruction-tuning dataset for object reasoning, and lack reflective validation of segmentation results, leading to erroneous mask predictions. To address these issues, in this paper, we propose a novel training-free Multi-Agent Recognition, Reasoning, and Reflection framework to achieve high-quality Reference Audio-Visual Segmentation, termed MAR3. Incorporating the sociological Delphi theory to achieve robust analysis, a Consensus Multimodal Recognition mechanism is proposed that enables LLM agents to explicitly recognize the difficulty of reference expressions and the dominant modality of multimodal cues. Based on our modality-dominant difficulty rule, we propose an adaptive Collaborative Object Reasoning strategy to reliably reason about the referred object. To further ensure precise mask prediction, we develop a Reflective Learning Segmentation mechanism, in which a check agent examines intermediate segmentation results and iteratively corrects the object text prompt of the segment agent. Experiments demonstrate that MAR3 achieves superior performance (69.2% in J&F) on the Ref-AVSBench dataset, outperforming SOTA by 3.4% absolutely.

Quanjun Zhang, Chunrong Fang, Haichuan Hu et al.

Automated program repair (APR) attempts to generate correct patches and has drawn wide attention from both academia and industry in the past decades. However, APR is continuously struggling with the patch overfitting issue due to the weak test suites. Thus, to address the overfitting problem, the community has proposed an increasing number of approaches to predict patch correctness (APCA approaches). Among them, locally deep learning approaches aimed at automatically match designs has been emerging strongly. Such approaches typically encode input code snippets into well-designed representations and build a binary model for correctness prediction. Despite being fundamental in reason about patch correctness, code representation has not been systematically investigated. To bridge this gap, we perform the first extensive study to evaluate the performance of different code representations on predicting patch correctness from more than 500 trained APCA models. The experimental results on 15 benchmarks with four categories and 11 classifiers show that the graph-based code representation which is ill-explored in the literature, consistently outperforms other representations, e.g., an average accuracy of 82.6% for CPG across three GNN models. Moreover, we demonstrate that such representations can achieve comparable or better performance for three different previous APCA approaches, e.g., filtering out 87.09% overfitting patches by TREETRAIN with AST. We further find that integrating sequence-based representation into heuristic-based representation is able to yield an average improvement of 13.5% on five metrics. Overall, our study highlights the potential and challenges of utilizing code representation to reason about patch correctness, thus increasing the usability of off-the-shelf APR tools and reducing the manual debugging effort of developers in practice.

Yuan Zhao, Lin Liu

Text-to-image diffusion models have achieved remarkable image quality, but they still struggle with complex, multiele ment prompts, and limited stylistic diversity. To address these limitations, we propose a Multi-Expert Planning and Gen eration Framework (MEPG) that synergistically integrates position- and style-aware large language models (LLMs) with spatial-semantic expert modules. The framework comprises two core components: (1) a Position-Style-Aware (PSA) module that utilizes a supervised fine-tuned LLM to decom pose input prompts into precise spatial coordinates and style encoded semantic instructions; and (2) a Multi-Expert Dif fusion (MED) module that implements cross-region genera tion through dynamic expert routing across both local regions and global areas. During the generation process for each lo cal region, specialized models (e.g., realism experts, styliza tion specialists) are selectively activated for each spatial par tition via attention-based gating mechanisms. The architec ture supports lightweight integration and replacement of ex pert models, providing strong extensibility. Additionally, an interactive interface enables real-time spatial layout editing and per-region style selection from a portfolio of experts. Ex periments show that MEPG significantly outperforms base line models with the same backbone in both image quality and style diversity.

Kai Yao, Zhaorui Tan, Penglei Gao et al.

The rapid growth of large language models (LLMs) with traditional centralized fine-tuning emerges as a key technique for adapting these models to domain-specific challenges, yielding privacy risks for both model and data owners. One promising solution, called offsite-tuning (OT), is proposed to address these challenges, where a weaker emulator is compressed from the original model and further fine-tuned with adapter to enhance privacy. However, the existing OT-based methods require high computational costs and lack theoretical analysis. This paper introduces a novel OT approach based on gradient-preserving compression, named GradOT. By analyzing the OT problem through the lens of optimization, we propose a method that selectively applies compression techniques such as rank compression and channel pruning, preserving the gradients of fine-tuned adapters while ensuring privacy. Extensive experiments demonstrate that our approach surpasses existing OT methods, both in terms of privacy protection and model performance. Our method provides a theoretical foundation for OT and offers a practical, training-free solution for offsite-tuning of large-scale LLMs.

Qi Li, Shaheer U. Saeed, Yuliang Huang et al.

Trackerless freehand ultrasound reconstruction aims to reconstruct 3D volumes from sequences of 2D ultrasound images without relying on external tracking systems. By eliminating the need for optical or electromagnetic trackers, this approach offers a low-cost, portable, and widely deployable alternative to more expensive volumetric ultrasound imaging systems, particularly valuable in resource-constrained clinical settings. However, predicting long-distance transformations and handling complex probe trajectories remain challenging. The TUS-REC2024 Challenge establishes the first benchmark for trackerless 3D freehand ultrasound reconstruction by providing a large publicly available dataset, along with a baseline model and a rigorous evaluation framework. By the submission deadline, the Challenge had attracted 43 registered teams, of which 6 teams submitted 21 valid dockerized solutions. The submitted methods span a wide range of approaches, including the state space model, the recurrent model, the registration-driven volume refinement, the attention mechanism, and the physics-informed model. This paper provides a comprehensive background introduction and literature review in the field, presents an overview of the challenge design and dataset, and offers a comparative analysis of submitted methods across multiple evaluation metrics. These analyses highlight both the progress and the current limitations of state-of-the-art approaches in this domain and provide insights for future research directions. All data and code are publicly available to facilitate ongoing development and reproducibility. As a live and evolving benchmark, it is designed to be continuously iterated and improved. The Challenge was held at MICCAI 2024 and is organised again at MICCAI 2025, reflecting its sustained commitment to advancing this field.

Zhi Qin, Qianhui Gui, Mouxiao Bian et al.

Medical imaging quality control (QC) is essential for accurate diagnosis, yet traditional QC methods remain labor-intensive and subjective. To address this challenge, in this study, we establish a standardized dataset and evaluation framework for medical imaging QC, systematically assessing large language models (LLMs) in image quality assessment and report standardization. Specifically, we first constructed and anonymized a dataset of 161 chest X-ray (CXR) radiographs and 219 CT reports for evaluation. Then, multiple LLMs, including Gemini 2.0-Flash, GPT-4o, and DeepSeek-R1, were evaluated based on recall, precision, and F1 score to detect technical errors and inconsistencies. Experimental results show that Gemini 2.0-Flash achieved a Macro F1 score of 90 in CXR tasks, demonstrating strong generalization but limited fine-grained performance. DeepSeek-R1 excelled in CT report auditing with a 62.23\% recall rate, outperforming other models. However, its distilled variants performed poorly, while InternLM2.5-7B-chat exhibited the highest additional discovery rate, indicating broader but less precise error detection. These findings highlight the potential of LLMs in medical imaging QC, with DeepSeek-R1 and Gemini 2.0-Flash demonstrating superior performance.

Patrick Lee, Alain Chirino Trujillo, Diana Cuevas Plancarte et al.

This study investigates the computational processing of euphemisms, a universal linguistic phenomenon, across multiple languages. We train a multilingual transformer model (XLM-RoBERTa) to disambiguate potentially euphemistic terms (PETs) in multilingual and cross-lingual settings. In line with current trends, we demonstrate that zero-shot learning across languages takes place. We also show cases where multilingual models perform better on the task compared to monolingual models by a statistically significant margin, indicating that multilingual data presents additional opportunities for models to learn about cross-lingual, computational properties of euphemisms. In a follow-up analysis, we focus on universal euphemistic "categories" such as death and bodily functions among others. We test to see whether cross-lingual data of the same domain is more important than within-language data of other domains to further understand the nature of the cross-lingual transfer.

Patrick Lee, Iyanuoluwa Shode, Alain Chirino Trujillo et al.

Transformers have been shown to work well for the task of English euphemism disambiguation, in which a potentially euphemistic term (PET) is classified as euphemistic or non-euphemistic in a particular context. In this study, we expand on the task in two ways. First, we annotate PETs for vagueness, a linguistic property associated with euphemisms, and find that transformers are generally better at classifying vague PETs, suggesting linguistic differences in the data that impact performance. Second, we present novel euphemism corpora in three different languages: Yoruba, Spanish, and Mandarin Chinese. We perform euphemism disambiguation experiments in each language using multilingual transformer models mBERT and XLM-RoBERTa, establishing preliminary results from which to launch future work.

Matthew Dowling, Yuan Zhao, Il Memming Park

Latent variable models have become instrumental in computational neuroscience for reasoning about neural computation. This has fostered the development of powerful offline algorithms for extracting latent neural trajectories from neural recordings. However, despite the potential of real time alternatives to give immediate feedback to experimentalists, and enhance experimental design, they have received markedly less attention. In this work, we introduce the exponential family variational Kalman filter (eVKF), an online recursive Bayesian method aimed at inferring latent trajectories while simultaneously learning the dynamical system generating them. eVKF works for arbitrary likelihoods and utilizes the constant base measure exponential family to model the latent state stochasticity. We derive a closed-form variational analogue to the predict step of the Kalman filter which leads to a provably tighter bound on the ELBO compared to another online variational method. We validate our method on synthetic and real-world data, and, notably, show that it achieves competitive performance

Luanxuan Hou, Jie Cao, Yuan Zhao et al.

We propose an augmented Parallel-Pyramid Net ($P^2~Net$) with feature refinement by dilated bottleneck and attention module. During data preprocessing, we proposed a differentiable auto data augmentation ($DA^2$) method. We formulate the problem of searching data augmentaion policy in a differentiable form, so that the optimal policy setting can be easily updated by back propagation during training. $DA^2$ improves the training efficiency. A parallel-pyramid structure is followed to compensate the information loss introduced by the network. We innovate two fusion structures, i.e. Parallel Fusion and Progressive Fusion, to process pyramid features from backbone network. Both fusion structures leverage the advantages of spatial information affluence at high resolution and semantic comprehension at low resolution effectively. We propose a refinement stage for the pyramid features to further boost the accuracy of our network. By introducing dilated bottleneck and attention module, we increase the receptive field for the features with limited complexity and tune the importance to different feature channels. To further refine the feature maps after completion of feature extraction stage, an Attention Module ($AM$) is defined to extract weighted features from different scale feature maps generated by the parallel-pyramid structure. Compared with the traditional up-sampling refining, $AM$ can better capture the relationship between channels. Experiments corroborate the effectiveness of our proposed method. Notably, our method achieves the best performance on the challenging MSCOCO and MPII datasets.

Diego M. Arribas, Yuan Zhao, Il Memming Park

The standard approach to fitting an autoregressive spike train model is to maximize the likelihood for one-step prediction. This maximum likelihood estimation (MLE) often leads to models that perform poorly when generating samples recursively for more than one time step. Moreover, the generated spike trains can fail to capture important features of the data and even show diverging firing rates. To alleviate this, we propose to directly minimize the divergence between neural recorded and model generated spike trains using spike train kernels. We develop a method that stochastically optimizes the maximum mean discrepancy induced by the kernel. Experiments performed on both real and synthetic neural data validate the proposed approach, showing that it leads to well-behaving models. Using different combinations of spike train kernels, we show that we can control the trade-off between different features which is critical for dealing with model-mismatch.

Matthew Dowling, Yuan Zhao, Il Memming Park

A fundamental problem in statistical neuroscience is to model how neurons encode information by analyzing electrophysiological recordings. A popular and widely-used approach is to fit the spike trains with an autoregressive point process model. These models are characterized by a set of convolutional temporal filters, whose subsequent analysis can help reveal how neurons encode stimuli, interact with each other, and process information. In practice a sufficiently rich but small ensemble of temporal basis functions needs to be chosen to parameterize the filters. However, obtaining a satisfactory fit often requires burdensome model selection and fine tuning the form of the basis functions and their temporal span. In this paper we propose a nonparametric approach for jointly inferring the filters and hyperparameters using the Gaussian process framework. Our method is computationally efficient taking advantage of the sparse variational approximation while being flexible and rich enough to characterize arbitrary filters in continuous time lag. Moreover, our method automatically learns the temporal span of the filter. For the particular application in neuroscience, we designed priors for stimulus and history filters useful for the spike trains. We compare and validate our method on simulated and real neural spike train data.

Vishwali Mhasawade, Yuan Zhao, Rumi Chunara

Research in population and public health focuses on the mechanisms between different cultural, social, and environmental factors and their effect on the health, of not just individuals, but communities as a whole. We present here a very brief introduction into research in these fields, as well as connections to existing machine learning work to help activate the machine learning community on such topics and highlight specific opportunities where machine learning, public and population health may synergize to better achieve health equity.

Luanxuan Hou, Jie Cao, Yuan Zhao et al.

The target of human pose estimation is to determine body part or joint locations of each person from an image. This is a challenging problems with wide applications. To address this issue, this paper proposes an augmented parallel-pyramid net with attention partial module and differentiable auto-data augmentation. Technically, a parallel pyramid structure is proposed to compensate the loss of information. We take the design of parallel structure for reverse compensation. Meanwhile, the overall computational complexity does not increase. We further define an Attention Partial Module (APM) operator to extract weighted features from different scale feature maps generated by the parallel pyramid structure. Compared with refining through upsampling operator, APM can better capture the relationship between channels. At last, we proposed a differentiable auto data augmentation method to further improve estimation accuracy. We define a new pose search space where the sequences of data augmentations are formulated as a trainable and operational CNN component. Experiments corroborate the effectiveness of our proposed method. Notably, our method achieves the top-1 accuracy on the challenging COCO keypoint benchmark and the state-of-the-art results on the MPII datasets.

Yuan Zhao, Jiasi Chen, Samet Oymak

Safety-critical applications require machine learning models that output accurate and calibrated probabilities. While uncalibrated deep networks are known to make over-confident predictions, it is unclear how model confidence is impacted by the variations in the data, such as label noise or class size. In this paper, we investigate the role of the dataset quality by studying the impact of dataset size and the label noise on the model confidence. We theoretically explain and experimentally demonstrate that, surprisingly, label noise in the training data leads to under-confident networks, while reduced dataset size leads to over-confident models. We then study the impact of dataset heterogeneity, where data quality varies across classes, on model confidence. We demonstrate that this leads to heterogenous confidence/accuracy behavior in the test data and is poorly handled by the standard calibration algorithms. To overcome this, we propose an intuitive heterogenous calibration technique and show that the proposed approach leads to improved calibration metrics (both average and worst-case errors) on the CIFAR datasets.

Yuan Zhao, Josue Nassar, Ian Jordan et al.

Nonlinear state-space models are powerful tools to describe dynamical structures in complex time series. In a streaming setting where data are processed one sample at a time, simultaneous inference of the state and its nonlinear dynamics has posed significant challenges in practice. We develop a novel online learning framework, leveraging variational inference and sequential Monte Carlo, which enables flexible and accurate Bayesian joint filtering. Our method provides an approximation of the filtering posterior which can be made arbitrarily close to the true filtering distribution for a wide class of dynamics models and observation models. Specifically, the proposed framework can efficiently approximate a posterior over the dynamics using sparse Gaussian processes, allowing for an interpretable model of the latent dynamics. Constant time complexity per sample makes our approach amenable to online learning scenarios and suitable for real-time applications.

Siyang Sun, Yingjie Yin, Xingang Wang et al.

Object detection plays an important role in various visual applications. However, the precision and speed of detector are usually contradictory. One main reason for fast detectors' precision reduction is that small objects are hard to be detected. To address this problem, we propose a multiple receptive field and small-object-focusing weakly-supervised segmentation network (MRFSWSnet) to achieve fast object detection. In MRFSWSnet, multiple receptive fields block (MRF) is used to pay attention to the object and its adjacent background's different spatial location with different weights to enhance the feature's discriminability. In addition, in order to improve the accuracy of small object detection, a small-object-focusing weakly-supervised segmentation module which only focuses on small object instead of all objects is integrated into the detection network for auxiliary training to improve the precision of small object detection. Extensive experiments show the effectiveness of our method on both PASCAL VOC and MS COCO detection datasets. In particular, with a lower resolution version of 300x300, MRFSWSnet achieves 80.9% mAP on VOC2007 test with an inference speed of 15 milliseconds per frame, which is the state-of-the-art detector among real-time detectors.

Yiwei Zhang, Chunbiao Zhu, Ge Li et al.

A fast and effective motion deblurring method has great application values in real life. This work presents an innovative approach in which a self-paced learning is combined with GAN to deblur image. First, We explain that a proper generator can be used as deep priors and point out that the solution for pixel-based loss is not same with the one for perception-based loss. By using these ideas as starting points, a Bi-Skip network is proposed to improve the generating ability and a bi-level loss is adopted to solve the problem that common conditions are non-identical. Second, considering that the complex motion blur will perturb the network in the training process, a self-paced mechanism is adopted to enhance the robustness of the network. Through extensive evaluations on both qualitative and quantitative criteria, it is demonstrated that our approach has a competitive advantage over state-of-the-art methods.

Wanxin Tian, Zixuan Wang, Haifeng Shen et al.

The performance of face detectors has been largely improved with the development of convolutional neural network. However, it remains challenging for face detectors to detect tiny, occluded or blurry faces. Besides, most face detectors can't locate face's position precisely and can't achieve high Intersection-over-Union (IoU) scores. We assume that problems inside are inadequate use of supervision information and imbalance between semantics and details at all level feature maps in CNN even with Feature Pyramid Networks (FPN). In this paper, we present a novel single-shot face detection network, named DF$^2$S$^2$ (Detection with Feature Fusion and Segmentation Supervision), which introduces a more effective feature fusion pyramid and a more efficient segmentation branch on ResNet-50 to handle mentioned problems. Specifically, inspired by FPN and SENet, we apply semantic information from higher-level feature maps as contextual cues to augment low-level feature maps via a spatial and channel-wise attention style, preventing details from being covered by too much semantics and making semantics and details complement each other. We further propose a semantic segmentation branch to best utilize detection supervision information meanwhile applying attention mechanism in a self-supervised manner. The segmentation branch is supervised by weak segmentation ground-truth (no extra annotation is required) in a hierarchical manner, deprecated in the inference time so it wouldn't compromise the inference speed. We evaluate our model on WIDER FACE dataset and achieved state-of-art results.

Guoqing Chao, Chengsheng Mao, Fei Wang et al.

ICU mortality risk prediction is a tough yet important task. On one hand, due to the complex temporal data collected, it is difficult to identify the effective features and interpret them easily; on the other hand, good prediction can help clinicians take timely actions to prevent the mortality. These correspond to the interpretability and accuracy problems. Most existing methods lack of the interpretability, but recently Subgraph Augmented Nonnegative Matrix Factorization (SANMF) has been successfully applied to time series data to provide a path to interpret the features well. Therefore, we adopted this approach as the backbone to analyze the patient data. One limitation of the raw SANMF method is its poor prediction ability due to its unsupervised nature. To deal with this problem, we proposed a supervised SANMF algorithm by integrating the logistic regression loss function into the NMF framework and solved it with an alternating optimization procedure. We used the simulation data to verify the effectiveness of this method, and then we applied it to ICU mortality risk prediction and demonstrated its superiority over other conventional supervised NMF methods.

Chengsheng Mao, Yuan Zhao, Mengxin Sun et al.

Privacy is a major concern in sharing human subject data to researchers for secondary analyses. A simple binary consent (opt-in or not) may significantly reduce the amount of sharable data, since many patients might only be concerned about a few sensitive medical conditions rather than the entire medical records. We propose event-level privacy protection, and develop a feature ablation method to protect event-level privacy in electronic medical records. Using a list of 13 sensitive diagnoses, we evaluate the feasibility and the efficacy of the proposed method. As feature ablation progresses, the identifiability of a sensitive medical condition decreases with varying speeds on different diseases. We find that these sensitive diagnoses can be divided into 3 categories: (1) 5 diseases have fast declining identifiability (AUC below 0.6 with less than 400 features excluded); (2) 7 diseases with progressively declining identifiability (AUC below 0.7 with between 200 and 700 features excluded); and (3) 1 disease with slowly declining identifiability (AUC above 0.7 with 1000 features excluded). The fact that the majority (12 out of 13) of the sensitive diseases fall into the first two categories suggests the potential of the proposed feature ablation method as a solution for event-level record privacy protection.

Yuan Zhao, Il Memming Park

New technologies for recording the activity of large neural populations during complex behavior provide exciting opportunities for investigating the neural computations that underlie perception, cognition, and decision-making. Nonlinear state space models provide an interpretable signal processing framework by combining an intuitive dynamical system with a probabilistic observation model, which can provide insights into neural dynamics, neural computation, and development of neural prosthetics and treatment through feedback control. It brings the challenge of learning both latent neural state and the underlying dynamical system because neither is known for neural systems a priori. We developed a flexible online learning framework for latent nonlinear state dynamics and filtered latent states. Using the stochastic gradient variational Bayes approach, our method jointly optimizes the parameters of the nonlinear dynamical system, the observation model, and the black-box recognition model. Unlike previous approaches, our framework can incorporate non-trivial distributions of observation noise and has constant time and space complexity. These features make our approach amenable to real-time applications and the potential to automate analysis and experimental design in ways that testably track and modify behavior using stimuli designed to influence learning.

Yuan Zhao, Il Memming Park

When governed by underlying low-dimensional dynamics, the interdependence of simultaneously recorded population of neurons can be explained by a small number of shared factors, or a low-dimensional trajectory. Recovering these latent trajectories, particularly from single-trial population recordings, may help us understand the dynamics that drive neural computation. However, due to the biophysical constraints and noise in the spike trains, inferring trajectories from data is a challenging statistical problem in general. Here, we propose a practical and efficient inference method, called the variational latent Gaussian process (vLGP). The vLGP combines a generative model with a history-dependent point process observation together with a smoothness prior on the latent trajectories. The vLGP improves upon earlier methods for recovering latent trajectories, which assume either observation models inappropriate for point processes or linear dynamics. We compare and validate vLGP on both simulated datasets and population recordings from the primary visual cortex. In the V1 dataset, we find that vLGP achieves substantially higher performance than previous methods for predicting omitted spike trains, as well as capturing both the toroidal topology of visual stimuli space, and the noise-correlation. These results show that vLGP is a robust method with a potential to reveal hidden neural dynamics from large-scale neural recordings.