Yiyang Wang

Lucheng Fu, Ye Yu, Yiyang Wang et al.

Large language models (LLMs) are highly sensitive to the prompts used to specify task objectives and behavioral constraints. Many recent prompt optimization methods iteratively rewrite prompts using LLM-generated feedback, but the resulting prompts often become longer, accumulate narrow sample-specific rules, and generalize poorly beyond the training distribution. We study this failure mode as prompt distributional overfitting and argue that it reflects a lack of representation control in discrete text-space optimization. We formalize this view through representational inefficiency, a dual-factor measure that decomposes prompt inefficiency into capacity cost and scope narrowness, attributing distributional prompt overfitting to their coupled growth during optimization. We propose TextReg, a regularization framework that realizes a soft-penalty objective through regularized textual gradients, combining Dual-Evidence Gradient Purification, Semantic Edit Regularization, and Regularization-Guided Prompt Update. Across multiple reasoning benchmarks, TextReg substantially improves out-of-distribution (OOD) generalization, with accuracy gains of up to +11.8% over TextGrad and +16.5% over REVOLVE.

Yiyang Wang, Moeiini Reilly, Britney Johnson et al.

Gardening is critical to support well-being, cultural continuity, and food autonomy, yet existing digital tools often provide generic advice that overlooks gardeners' skills, local ecologies, seasons, and cultural contexts. We introduce CultivAgents, a relationship-centered multi-agent system for personalized, socio-culturally grounded gardening support. Grounded in ethics of care, CultivAgents coordinates multiple specialized agents: an Experience Agent that adapts guidance to users' skill levels, an Environmental Agent that grounds advice in local and seasonal conditions, and an Ethnobotanical Agent that connects plants to cultural knowledge and histories. We evaluated CultivAgents through a three-phase mixed-methods study with domain experts (n=3), HCI researchers (n=7), and community gardeners (n=5), analyzing expert feedback, pre/post surveys, and participatory design activities. Results suggest that CultivAgents helped gardeners translate interest into situated action: community gardeners reported increased confidence (3.00 to 3.60), motivation (4.00 to 4.40), and trust in acting on AI advice (3.20 to 4.00). Participants valued hyperlocal ecological guidance and complementary agent perspectives, while also identifying limits in cultural specificity, ecological grounding, and agent coordination. The work advances relationship-centered AI, offering design implications for multi-agent systems that support food sovereignty, community resilience, and cultural preservation.

Naiyao Wang, Tongbang Jiang, Ye Wang et al.

Intelligent maritime, as an essential component of smart ocean construction, deeply integrates advanced artificial intelligence technology and data analysis methods, which covers multiple aspects such as smart vessels, route optimization, safe navigation, aiming to enhance the efficiency of ocean resource utilization and the intelligence of transportation networks. However, the complex and dynamic maritime environment, along with diverse and heterogeneous large-scale data sources, present challenges for real-time decision-making in intelligent maritime. In this paper, We propose KUNPENG, the first-ever embodied large model for intelligent maritime in the smart ocean construction, which consists of six systems. The model perceives multi-source heterogeneous data for the cognition of environmental interaction and make autonomous decision strategies, which are used for intelligent vessels to perform navigation behaviors under safety and emergency guarantees and continuously optimize power to achieve embodied intelligence in maritime. In comprehensive maritime task evaluations, KUNPENG has demonstrated excellent performance.

Yiyang Wang, Yiqiao Jin, Alex Cabral et al.

Multi-agent systems (MAS) have recently emerged as promising socio-collaborative companions for emotional and cognitive support. However, these systems frequently suffer from persona collapse--where agents revert to generic, homogenized assistant behaviors--and social sycophancy, which produces redundant, non-constructive dialogue. We propose MASCOT, a generalizable framework for multi-perspective socio-collaborative companions. MASCOT introduces a novel bi-level optimization strategy to harmonize individual and collective behaviors: 1) Persona-Aware Behavioral Alignment, an RLAIF-driven pipeline that finetunes individual agents for strict persona fidelity to prevent identity loss; and 2) Collaborative Dialogue Optimization, a meta-policy guided by group-level rewards to ensure diverse and productive discourse. Extensive evaluations across psychological support and workplace domains demonstrate that MASCOT significantly outperforms state-of-the-art baselines, achieving improvements of up to +14.1 in Persona Consistency and +10.6 in Social Contribution. Our framework provides a practical roadmap for engineering the next generation of socially intelligent multi-agent systems.

Yiyang Wang, Chen Chen, Tica Lin et al.

Social presence is central to the enjoyment of watching content together, yet modern media consumption is increasingly solitary. We investigate whether multi-agent conversational AI systems can recreate the dynamics of shared viewing experiences across diverse content types. We present CompanionCast, a general framework for orchestrating multiple role-specialized AI agents that respond to video content using multimodal inputs, speech synthesis, and spatial audio. Distinctly, CompanionCast integrates an LLM-as-a-Judge module that iteratively scores and refines conversations across five dimensions (relevance, authenticity, engagement, diversity, personality consistency). We validate this framework through sports viewing, a domain with rich dynamics and strong social traditions, where a pilot study with soccer fans suggests that multi-agent interaction improves perceived social presence compared to solo viewing. We contribute: (1) a generalizable framework for orchestrating multi-agent conversations around multimodal video content, (2) a novel evaluator-agent pipeline for conversation quality control, and (3) exploratory evidence of increased social presence in AI-mediated co-viewing. We discuss challenges and future directions for applying this approach to diverse viewing contexts including entertainment, education, and collaborative watching experiences.

Adrian Manchado, Tanner Cellio, Jonathan Keane et al.

Sport analysis is crucial for team performance since it provides actionable data that can inform coaching decisions, improve player performance, and enhance team strategies. To analyze more complex features from game footage, a computer vision model can be used to identify and track key entities from the field. We propose the use of an object detection and tracking system to predict player positioning throughout the game. To translate this to positioning in relation to the field dimensions, we use a point prediction model to identify key points on the field and combine these with known field dimensions to extract actual distances. For the player-identification model, object detection models like YOLO and Faster R-CNN are evaluated on the accuracy of our custom video footage using multiple different evaluation metrics. The goal is to identify the best model for object identification to obtain the most accurate results when paired with SAM2 (Segment Anything Model 2) for segmentation and tracking. For the key point detection model, we use a CNN model to find consistent locations in the soccer field. Through homography, the positions of points and objects in the camera perspective will be transformed to a real-ground perspective. The segmented player masks from SAM2 are transformed from camera perspective to real-world field coordinates through homography, regardless of camera angle or movement. The transformed real-world coordinates can be used to calculate valuable tactical insights including player speed, distance covered, positioning heatmaps, and more complex team statistics, providing coaches and players with actionable performance data previously unavailable from standard video analysis.

Yiyang Wang, Qijia Zhou, Shengyuan Deng et al.

By integrating physics-informed neural network (PINN) techniques with domain decomposition method, a deep domain decomposition method is presented for solving elliptic variational inequality problems. Based on the Ritz variation method, the elliptic variational inequality problem is firstly reformulated as an optimization problem, and then the subproblem in each subdomain is solved by using the Ritz-PINN method, which the parameters in the network are updated by the Adam optimizer, and the residual-adaptive training by introducing a residual-adaptive dataset update strategy to gradually guide the model to learn more complex regions. Additionally, the impact of overlapping regions on the performance of the new algorithm is explored. Numerical results demonstrate the effectiveness of the proposed algorithm, the mean square error can be reached 1.0e-07, and the number of iterations is independent of grid length h under uniform overlap conditions.

Yiyang Wang, Xi Chen, Xiaogang Xu et al.

Recent advancements adopt online reinforcement learning (RL) from LLMs to text-to-image rectified flow diffusion models for reward alignment. The use of group-level rewards successfully aligns the model with the targeted reward. However, it faces challenges including low efficiency, dependency on stochastic samplers, and reward hacking. The problem is that rectified flow models are fundamentally different from LLMs: 1) For efficiency, online image sampling takes much more time and dominates the time of training. 2) For stochasticity, rectified flow is deterministic once the initial noise is fixed. Aiming at these problems and inspired by the effects of group-level rewards from LLMs, we design Group-level Direct Reward Optimization (GDRO). GDRO is a new post-training paradigm for group-level reward alignment that combines the characteristics of rectified flow models. Through rigorous theoretical analysis, we point out that GDRO supports full offline training that saves the large time cost for image rollout sampling. Also, it is diffusion-sampler-independent, which eliminates the need for the ODE-to-SDE approximation to obtain stochasticity. We also empirically study the reward hacking trap that may mislead the evaluation, and involve this factor in the evaluation using a corrected score that not only considers the original evaluation reward but also the trend of reward hacking. Extensive experiments demonstrate that GDRO effectively and efficiently improves the reward score of the diffusion model through group-wise offline optimization across the OCR and GenEval tasks, while demonstrating strong stability and robustness in mitigating reward hacking.

Yiqiao Jin, Qinlin Zhao, Yiyang Wang et al.

Peer review is fundamental to the integrity and advancement of scientific publication. Traditional methods of peer review analyses often rely on exploration and statistics of existing peer review data, which do not adequately address the multivariate nature of the process, account for the latent variables, and are further constrained by privacy concerns due to the sensitive nature of the data. We introduce AgentReview, the first large language model (LLM) based peer review simulation framework, which effectively disentangles the impacts of multiple latent factors and addresses the privacy issue. Our study reveals significant insights, including a notable 37.1% variation in paper decisions due to reviewers' biases, supported by sociological theories such as the social influence theory, altruism fatigue, and authority bias. We believe that this study could offer valuable insights to improve the design of peer review mechanisms. Our code is available at https://github.com/Ahren09/AgentReview.

Yiqiao Jin, Yijia Xiao, Yiyang Wang et al.

Understanding the creation, evolution, and dissemination of scientific knowledge is crucial for bridging diverse subject areas and addressing complex global challenges such as pandemics, climate change, and ethical AI. Scientometrics, the quantitative and qualitative study of scientific literature, provides valuable insights into these processes. We introduce SciEvo, a longitudinal scientometric dataset with over two million academic publications, providing comprehensive contents information and citation graphs to support cross-disciplinary analyses. SciEvo is easy to use and available across platforms, including GitHub, Kaggle, and HuggingFace. Using SciEvo, we conduct a temporal study spanning over 30 years to explore key questions in scientometrics: the evolution of academic terminology, citation patterns, and interdisciplinary knowledge exchange. Our findings reveal critical insights, such as disparities in epistemic cultures, knowledge production modes, and citation practices. For example, rapidly developing, application-driven fields like LLMs exhibit significantly shorter citation age (2.48 years) compared to traditional theoretical disciplines like oral history (9.71 years). Our data and analytic tools can be accessed at https://github.com/Ahren09/SciEvo.

Yiqiao Jin, Yiyang Wang, Lucheng Fu et al.

Self-distillation (SD) offers a promising path for adapting large language models (LLMs) without relying on stronger external teachers. However, SD in autoregressive LLMs remains challenging because self-generated trajectories are free-form, correctness is task-dependent, and plausible rationales can still provide unstable or unreliable supervision. Existing methods mainly examine isolated design choices, leaving their effectiveness, roles, and interactions unclear. In this paper, we propose UniSD, a unified framework to systematically study self-distillation. UniSD integrates complementary mechanisms that address supervision reliability, representation alignment, and training stability, including multi-teacher agreement, EMA teacher stabilization, token-level contrastive learning, feature matching, and divergence clipping. Across six benchmarks and six models from three model families, UniSD reveals when self-distillation improves over static imitation, which components drive the gains, and how these components interact across tasks. Guided by these insights, we construct UniSDfull, an integrated pipeline that combines complementary components and achieves the strongest overall performance, improving over the base model by +5.4 points and the strongest baseline by +2.8 points. Extensive evaluation highlights self-distillation as a practical and steerable approach for efficient LLM adaptation without stronger external teachers.

Qijia Zhou, Yiyang Wang, Shengyuan Deng et al.

Variational inequalities are widely applied in mechanical engineering, fluid penetration, transportation, and other fields. In this paper, a Deep Ritz method based on Physics-Informed Neural Networks (PINNs) is proposed to enhance the accuracy and efficiency of solving elliptic variational inequalities. The Ritz variational method is firstly utilized to transform the variational inequality problem into an optimization problem. Then Bayesian optimization is employed to tune the weights of the loss function, and a residual-based adaptive dataset update strategy is introduced to improve the convergence and accuracy of the model. Numerical experiments show that the proposed method can effectively approximate the analytical solution.

Sihui Ji, Hao Luo, Xi Chen et al.

We present LayerFlow, a unified solution for layer-aware video generation. Given per-layer prompts, LayerFlow generates videos for the transparent foreground, clean background, and blended scene. It also supports versatile variants like decomposing a blended video or generating the background for the given foreground and vice versa. Starting from a text-to-video diffusion transformer, we organize the videos for different layers as sub-clips, and leverage layer embeddings to distinguish each clip and the corresponding layer-wise prompts. In this way, we seamlessly support the aforementioned variants in one unified framework. For the lack of high-quality layer-wise training videos, we design a multi-stage training strategy to accommodate static images with high-quality layer annotations. Specifically, we first train the model with low-quality video data. Then, we tune a motion LoRA to make the model compatible with static frames. Afterward, we train the content LoRA on the mixture of image data with high-quality layered images along with copy-pasted video data. During inference, we remove the motion LoRA thus generating smooth videos with desired layers.

Yiyang Wang, Xi Chen, Xiaogang Xu et al.

The depth-of-field (DoF) effect, which introduces aesthetically pleasing blur, enhances photographic quality but is fixed and difficult to modify once the image has been created. This becomes problematic when the applied blur is undesirable~(e.g., the subject is out of focus). To address this, we propose DiffCamera, a model that enables flexible refocusing of a created image conditioned on an arbitrary new focus point and a blur level. Specifically, we design a diffusion transformer framework for refocusing learning. However, the training requires pairs of data with different focus planes and bokeh levels in the same scene, which are hard to acquire. To overcome this limitation, we develop a simulation-based pipeline to generate large-scale image pairs with varying focus planes and bokeh levels. With the simulated data, we find that training with only a vanilla diffusion objective often leads to incorrect DoF behaviors due to the complexity of the task. This requires a stronger constraint during training. Inspired by the photographic principle that photos of different focus planes can be linearly blended into a multi-focus image, we propose a stacking constraint during training to enforce precise DoF manipulation. This constraint enhances model training by imposing physically grounded refocusing behavior that the focusing results should be faithfully aligned with the scene structure and the camera conditions so that they can be combined into the correct multi-focus image. We also construct a benchmark to evaluate the effectiveness of our refocusing model. Extensive experiments demonstrate that DiffCamera supports stable refocusing across a wide range of scenes, providing unprecedented control over DoF adjustments for photography and generative AI applications.

Yiyang Wang, Xi Chen, Xiaogang Xu et al.

In spite of recent progress, image diffusion models still produce artifacts. A common solution is to leverage the feedback provided by quality assessment systems or human annotators to optimize the model, where images are generally rated in their entirety. In this work, we believe problem-solving starts with identification, yielding the request that the model should be aware of not just the presence of defects in an image, but their specific locations. Motivated by this, we propose DiffDoctor, a two-stage pipeline to assist image diffusion models in generating fewer artifacts. Concretely, the first stage targets developing a robust artifact detector, for which we collect a dataset of over 1M flawed synthesized images and set up an efficient human-in-the-loop annotation process, incorporating a carefully designed class-balance strategy. The learned artifact detector is then involved in the second stage to optimize the diffusion model by providing pixel-level feedback. Extensive experiments on text-to-image diffusion models demonstrate the effectiveness of our artifact detector as well as the soundness of our diagnose-then-treat design.

Sihui Ji, Yiyang Wang, Xi Chen et al.

We present FashionComposer for compositional fashion image generation. Unlike previous methods, FashionComposer is highly flexible. It takes multi-modal input (i.e., text prompt, parametric human model, garment image, and face image) and supports personalizing the appearance, pose, and figure of the human and assigning multiple garments in one pass. To achieve this, we first develop a universal framework capable of handling diverse input modalities. We construct scaled training data to enhance the model's robust compositional capabilities. To accommodate multiple reference images (garments and faces) seamlessly, we organize these references in a single image as an "asset library" and employ a reference UNet to extract appearance features. To inject the appearance features into the correct pixels in the generated result, we propose subject-binding attention. It binds the appearance features from different "assets" with the corresponding text features. In this way, the model could understand each asset according to their semantics, supporting arbitrary numbers and types of reference images. As a comprehensive solution, FashionComposer also supports many other applications like human album generation, diverse virtual try-on tasks, etc.

Xi Chen, Yutong Feng, Mengting Chen et al.

Image editing serves as a practical yet challenging task considering the diverse demands from users, where one of the hardest parts is to precisely describe how the edited image should look like. In this work, we present a new form of editing, termed imitative editing, to help users exercise their creativity more conveniently. Concretely, to edit an image region of interest, users are free to directly draw inspiration from some in-the-wild references (e.g., some relative pictures come across online), without having to cope with the fit between the reference and the source. Such a design requires the system to automatically figure out what to expect from the reference to perform the editing. For this purpose, we propose a generative training framework, dubbed MimicBrush, which randomly selects two frames from a video clip, masks some regions of one frame, and learns to recover the masked regions using the information from the other frame. That way, our model, developed from a diffusion prior, is able to capture the semantic correspondence between separate images in a self-supervised manner. We experimentally show the effectiveness of our method under various test cases as well as its superiority over existing alternatives. We also construct a benchmark to facilitate further research.

Yiyang Wang, Neda Masoud

We extend the adversarial/non-stochastic multi-play multi-armed bandit (MPMAB) to the case where the number of arms to play is variable. The work is motivated by the fact that the resources allocated to scan different critical locations in an interconnected transportation system change dynamically over time and depending on the environment. By modeling the malicious hacker and the intrusion monitoring system as the attacker and the defender, respectively, we formulate the problem for the two players as a sequential pursuit-evasion game. We derive the condition under which a Nash equilibrium of the strategic game exists. For the defender side, we provide an exponential-weighted based algorithm with sublinear pseudo-regret. We further extend our model to heterogeneous rewards for both players, and obtain lower and upper bounds on the average reward for the attacker. We provide numerical experiments to demonstrate the effectiveness of a variable-arm play.

Yiyang Wang, Neda Masoud, Anahita Khojandi

In this paper we propose a novel observer-based method for anomaly detection in connected and automated vehicles (CAVs). The proposed method utilizes an augmented extended Kalman filter (AEKF) to smooth sensor readings of a CAV based on a nonlinear car-following motion model with time delay, where the leading vehicle's trajectory is used by the subject vehicle to detect sensor anomalies. We use the classic $χ^2$ fault detector in conjunction with the proposed AEKF for anomaly detection. To make the proposed model more suitable for real-world applications, we consider a stochastic communication time delay in the car-following model. Our experiments conducted on real-world connected vehicle data indicate that the AEKF with $χ^2$-detector can achieve a high anomaly detection performance.

Yiyang Wang, Neda Masoud, Anahita Khojandi

In this paper we propose a novel observer-based method to improve the safety and security of connected and automated vehicle (CAV) transportation. The proposed method combines model-based signal filtering and anomaly detection methods. Specifically, we use adaptive extended Kalman filter (AEKF) to smooth sensor readings of a CAV based on a nonlinear car-following motion model. Under the assumption of a car-following model, the subject vehicle utilizes its leading vehicle's information to detect sensor anomalies by employing previously-trained One Class Support Vector Machine (OCSVM) models. This approach allows the AEKF to estimate the state of a vehicle not only based on the vehicle's location and speed, but also by taking into account the state of the surrounding traffic. A communication time delay factor is considered in the car-following model to make it more suitable for real-world applications. Our experiments show that compared with the AEKF with a traditional $χ^2$-detector, our proposed method achieves a better anomaly detection performance. We also demonstrate that a larger time delay factor has a negative impact on the overall detection performance.

Yiyang Wang, Risheng Liu, Long Ma et al.

We in this paper propose a realizable framework TECU, which embeds task-specific strategies into update schemes of coordinate descent, for optimizing multivariate non-convex problems with coupled objective functions. On one hand, TECU is capable of improving algorithm efficiencies through embedding productive numerical algorithms, for optimizing univariate sub-problems with nice properties. From the other side, it also augments probabilities to receive desired results, by embedding advanced techniques in optimizations of realistic tasks. Integrating both numerical algorithms and advanced techniques together, TECU is proposed in a unified framework for solving a class of non-convex problems. Although the task embedded strategies bring inaccuracies in sub-problem optimizations, we provide a realizable criterion to control the errors, meanwhile, to ensure robust performances with rigid theoretical analyses. By respectively embedding ADMM and a residual-type CNN in our algorithm framework, the experimental results verify both efficiency and effectiveness of embedding task-oriented strategies in coordinate descent for solving practical problems.

Risheng Liu, Long Ma, Yiyang Wang et al.

Enhancing visual qualities of images plays very important roles in various vision and learning applications. In the past few years, both knowledge-driven maximum a posterior (MAP) with prior modelings and fully data-dependent convolutional neural network (CNN) techniques have been investigated to address specific enhancement tasks. In this paper, by exploiting the advantages of these two types of mechanisms within a complementary propagation perspective, we propose a unified framework, named deep prior ensemble (DPE), for solving various image enhancement tasks. Specifically, we first establish the basic propagation scheme based on the fundamental image modeling cues and then introduce residual CNNs to help predicting the propagation direction at each stage. By designing prior projections to perform feedback control, we theoretically prove that even with experience-inspired CNNs, DPE is definitely converged and the output will always satisfy our fundamental task constraints. The main advantage against conventional optimization-based MAP approaches is that our descent directions are learned from collected training data, thus are much more robust to unwanted local minimums. While, compared with existing CNN type networks, which are often designed in heuristic manners without theoretical guarantees, DPE is able to gain advantages from rich task cues investigated on the bases of domain knowledges. Therefore, DPE actually provides a generic ensemble methodology to integrate both knowledge and data-based cues for different image enhancement tasks. More importantly, our theoretical investigations verify that the feedforward propagations of DPE are properly controlled toward our desired solution. Experimental results demonstrate that the proposed DPE outperforms state-of-the-arts on a variety of image enhancement tasks in terms of both quantitative measure and visual perception quality.

Yiyang Wang, Risheng Liu, Xiaoliang Song et al.

In recent years, numerous vision and learning tasks have been (re)formulated as nonconvex and nonsmooth programmings(NNPs). Although some algorithms have been proposed for particular problems, designing fast and flexible optimization schemes with theoretical guarantee is a challenging task for general NNPs. It has been investigated that performing inexact inner iterations often benefit to special applications case by case, but their convergence behaviors are still unclear. Motivated by these practical experiences, this paper designs a novel algorithmic framework, named inexact proximal alternating direction method (IPAD) for solving general NNPs. We demonstrate that any numerical algorithms can be incorporated into IPAD for solving subproblems and the convergence of the resulting hybrid schemes can be consistently guaranteed by a series of simple error conditions. Beyond the guarantee in theory, numerical experiments on both synthesized and real-world data further demonstrate the superiority and flexibility of our IPAD framework for practical use.

Risheng Liu, Jing Wang, Yiyang Wang et al.

In this paper, we propose a novel sparse coding and counting method under Bayesian framwork for visual tracking. In contrast to existing methods, the proposed method employs the combination of L0 and L1 norm to regularize the linear coefficients of incrementally updated linear basis. The sparsity constraint enables the tracker to effectively handle difficult challenges, such as occlusion or image corruption. To achieve realtime processing, we propose a fast and efficient numerical algorithm for solving the proposed model. Although it is an NP-hard problem, the proposed accelerated proximal gradient (APG) approach is guaranteed to converge to a solution quickly. Besides, we provide a closed solution of combining L0 and L1 regularized representation to obtain better sparsity. Experimental results on challenging video sequences demonstrate that the proposed method achieves state-of-the-art results both in accuracy and speed.