arXiv subjects

Yuhang Wang

Publications and source records attributed to Yuhang Wang.

At least 73 records · Page 4Linked to original sources

arXiv2025

Learning from Risk: LLM-Guided Generation of Safety-Critical Scenarios with Prior Knowledge

Autonomous driving faces critical challenges in rare long-tail events and complex multi-agent interactions, which are scarce in real-world data yet essential for robust safety validation. This paper presents a high-fidelity scenario generation framework that integrates a conditional variational autoencoder (CVAE) with a large language model (LLM). The CVAE encodes historical trajectories and map information from large-scale naturalistic datasets to learn latent traffic structures, enabling the generation of physically consistent base scenarios. Building on this, the LLM acts as an adversarial reasoning engine, parsing unstructured scene descriptions into domain-specific loss functions and dynamically guiding scenario generation across varying risk levels. This knowledge-driven optimization balances realism with controllability, ensuring that generated scenarios remain both plausible and risk-sensitive. Extensive experiments in CARLA and SMARTS demonstrate that our framework substantially increases the coverage of high-risk and long-tail events, improves consistency between simulated and real-world traffic distributions, and exposes autonomous driving systems to interactions that are significantly more challenging than those produced by existing rule- or data-driven methods. These results establish a new pathway for safety validation, enabling principled stress-testing of autonomous systems under rare but consequential events.

cs.LG↗

arXiv2025

Reasoning Shapes Alignment: Investigating Cultural Alignment in Large Reasoning Models with Cultural Norms

The advanced reasoning capabilities of Large Reasoning Models enable them to thoroughly understand and apply safety policies through deliberate thought processes, thereby improving the models' safety. Beyond safety, these models must also be able to reflect the diverse range of human values across various cultures. This paper presents the Cultural Norm-based Cultural Alignment (CNCA) framework, which enables models to leverage their powerful reasoning ability to align with cultural norms. Specifically, we propose three methods to automatically mine cultural norms from limited survey data and explore ways to effectively utilize these norms for improving cultural alignment. Two alignment paradigms are examined: an in-context alignment method, where cultural norms are explicitly integrated into the user context, and a fine-tuning-based method, which internalizes norms through enhanced Chain-of-Thought training data. Comprehensive experiments demonstrate the effectiveness of these methods, highlighting that models with stronger reasoning capabilities benefit more from cultural norm mining and utilization. Our findings emphasize the potential for reasoning models to better reflect diverse human values through culturally informed alignment strategies.

cs.AI↗

arXiv2025

Beyond Pipelines: A Survey of the Paradigm Shift toward Model-Native Agentic AI

The rapid evolution of agentic AI marks a new phase in artificial intelligence, where Large Language Models (LLMs) no longer merely respond but act, reason, and adapt. This survey traces the paradigm shift in building agentic AI: from Pipeline-based systems, where planning, tool use, and memory are orchestrated by external logic, to the emerging Model-native paradigm, where these capabilities are internalized within the model's parameters. We first position Reinforcement Learning (RL) as the algorithmic engine enabling this paradigm shift. By reframing learning from imitating static data to outcome-driven exploration, RL underpins a unified solution of LLM + RL + Task across language, vision and embodied domains. Building on this, the survey systematically reviews how each capability -- Planning, Tool use, and Memory -- has evolved from externally scripted modules to end-to-end learned behaviors. Furthermore, it examines how this paradigm shift has reshaped major agent applications, specifically the Deep Research agent emphasizing long-horizon reasoning and the GUI agent emphasizing embodied interaction. We conclude by discussing the continued internalization of agentic capabilities like Multi-agent collaboration and Reflection, alongside the evolving roles of the system and model layers in future agentic AI. Together, these developments outline a coherent trajectory toward model-native agentic AI as an integrated learning and interaction framework, marking the transition from constructing systems that apply intelligence to developing models that grow intelligence through experience.

cs.AI↗

arXiv2025

InvDec: Inverted Decoder for Multivariate Time Series Forecasting with Separated Temporal and Variate Modeling

Multivariate time series forecasting requires simultaneously modeling temporal patterns and cross-variate dependencies. Channel-independent methods such as PatchTST excel at temporal modeling but ignore variable correlations, while pure variate-attention approaches such as iTransformer sacrifice temporal encoding. We proposeInvDec (Inverted Decoder), a hybrid architecture that achieves principled separation between temporal encoding and variate-level decoding. InvDec combines a patch-based temporal encoder with an inverted decoder operating on the variate dimension through variate-wise self-attention. We introduce delayed variate embeddings that enrich variable-specific representations only after temporal encoding, preserving temporal feature integrity. An adaptive residual fusion mechanism dynamically balances temporal and variate information across datasets of varying dimensions. Instantiating InvDec with PatchTST yields InvDec-PatchTST. Extensive experiments on seven benchmarks demonstrate significant gains on high-dimensional datasets: 20.9% MSE reduction on Electricity (321 variables), 4.3% improvement on Weather, and 2.7% gain on Traffic compared to PatchTST, while maintaining competitive performance on low-dimensional ETT datasets. Ablation studies validate each component, and analysis reveals that InvDec's advantage grows with dataset dimensionality, confirming that cross-variate modeling becomes critical as the number of variables increases.

cs.LG↗

arXiv2025

Fire-EnSF: Wildfire Spread Data Assimilation using Ensemble Score Filter

As wildfires become increasingly destructive and expensive to control, effective management of active wildfires requires accurate, real-time fire spread predictions. To enhance the forecasting accuracy of active fires, data assimilation plays a vital role by integrating observations (such as remote-sensing data) and fire predictions generated from numerical models. This paper provides a comprehensive investigation on the application of a recently proposed diffusion-model-based filtering algorithm -- the Ensemble Score Filter (EnSF) -- to the data assimilation problem for real-time active wildfire spread predictions. Leveraging a score-based generative diffusion model, EnSF has been shown to have superior accuracy for high-dimensional nonlinear filtering problems, making it an ideal candidate for the filtering problems of wildfire spread models. Technical details are provided, and our numerical investigations demonstrate that EnSF provides superior accuracy, stability, and computational efficiency, establishing it as a robust and practical method for wildfire data assimilation. Our code has been made publicly available.

cs.LG↗

arXiv2025

DM1: MeanFlow with Dispersive Regularization for 1-Step Robotic Manipulation

The ability to learn multi-modal action distributions is indispensable for robotic manipulation policies to perform precise and robust control. Flow-based generative models have recently emerged as a promising solution to learning distributions of actions, offering one-step action generation and thus achieving much higher sampling efficiency compared to diffusion-based methods. However, existing flow-based policies suffer from representation collapse, the inability to distinguish similar visual representations, leading to failures in precise manipulation tasks. We propose DM1 (MeanFlow with Dispersive Regularization for One-Step Robotic Manipulation), a novel flow matching framework that integrates dispersive regularization into MeanFlow to prevent collapse while maintaining one-step efficiency. DM1 employs multiple dispersive regularization variants across different intermediate embedding layers, encouraging diverse representations across training batches without introducing additional network modules or specialized training procedures. Experiments on RoboMimic benchmarks show that DM1 achieves 20-40 times faster inference (0.07s vs. 2-3.5s) and improves success rates by 10-20 percentage points, with the Lift task reaching 99% success over 85% of the baseline. Real-robot deployment on a Franka Panda further validates that DM1 transfers effectively from simulation to the physical world. To the best of our knowledge, this is the first work to leverage representation regularization to enable flow-based policies to achieve strong performance in robotic manipulation, establishing a simple yet powerful approach for efficient and robust manipulation.

cs.RO↗

arXiv2025

Don't Settle Too Early: Self-Reflective Remasking for Diffusion Language Models

Mask-based Diffusion Language Models (DLMs) struggle to revise incorrect tokens: once a token is generated, it typically remains fixed. The key challenge is to identify potential errors in the inputs. In this paper, we propose \emph{\underline{Rem}asking-\underline{e}nabled \underline{Di}ffusion Language Model (RemeDi}, a mask-based DLM that introduces \emph{remasking} as another fundamental mechanism, enabling more flexible text refinement in diffusion-based text generation. To achieve this, RemeDi jointly predicts token distributions and per-token confidence scores at each step. The confidence scores determine which tokens to be unmasked after the current step, allowing the model to identify tokens with low quality and remask them. These remasked tokens can be resampled with richer context in subsequent steps. We design a remask-aware pipeline to train this ability, including supervised fine-tuning which teaches the model to detect and remask incorrect tokens in addition to predict mask tokens, and reinforcement learning which optimizes full generation trajectories toward higher rewards. Experiments show that RemeDi achieves the state-of-the-art results among open-source DLMs on multiple datasets.

cs.CL↗

arXiv2025

C-Evolve: Consensus-based Evolution for Prompt Groups

Prompt evolution algorithms offer a powerful paradigm for enhancing AI systems based on closed-source models, while few work explores whether aggregating results from multiple prompts to reach a consensus can further advance the system capability boundary. In this paper, we introduce Consensus-Evolve (C-Evolve), an evolutionary algorithm that discovers a group of prompts whose aggregated outputs after majority voting achieve optimal performance. More specifically, C-Evolve employs an island-based evolutionary algorithm to maintain population diversity, and prompts from distinct islands are selected to form groups to aggregate their outputs. The key difference from single individual evolution is a voting score, which evaluates each individual prompt's contribution within groups. We take this as the fitness score for evolution instead of individual performance. Consequently, C-Evolve is more likely to produce and maintain prompts with higher potential to form a high-performing group and eliminate low-performing ones, gradually improving the group performance after reaching consensus. Our method achieves state-of-the-art performance across a wide range of tasks, including both open-ended tasks like HotpotQA and closed-ended tasks like MATH. On Qwen3-8B, C-Evolve achieves 70.67% on HotpotQA and 43.88% on IFBench, which are 4.95% and 2.73% higher than GEPA, respectively. For GPT-4.1-mini, the accuracy on IFBench is further improved to 47.96% and reaches 95.33% in the MATH benchmark. These results demonstrate the C-Evolve's competitive performance.

cs.CL↗

arXiv2025

Semantic-aware Graph-guided Behavior Sequences Generation with Large Language Models for Smart Homes

As smart homes become increasingly prevalent, intelligent models are widely used for tasks such as anomaly detection and behavior prediction. These models are typically trained on static datasets, making them brittle to behavioral drift caused by seasonal changes, lifestyle shifts, or evolving routines. However, collecting new behavior data for retraining is often impractical due to its slow pace, high cost, and privacy concerns. In this paper, we propose SmartGen, an LLM-based framework that synthesizes context-aware user behavior data to support continual adaptation of downstream smart home models. SmartGen consists of four key components. First, we design a Time and Semantic-aware Split module to divide long behavior sequences into manageable, semantically coherent subsequences under dual time-span constraints. Second, we propose Semantic-aware Sequence Compression to reduce input length while preserving representative semantics by clustering behavior mapping in latent space. Third, we introduce Graph-guided Sequence Synthesis, which constructs a behavior relationship graph and encodes frequent transitions into prompts, guiding the LLM to generate data aligned with contextual changes while retaining core behavior patterns. Finally, we design a Two-stage Outlier Filter to identify and remove implausible or semantically inconsistent outputs, aiming to improve the factual coherence and behavioral validity of the generated sequences. Experiments on three real-world datasets demonstrate that SmartGen significantly enhances model performance on anomaly detection and behavior prediction tasks under behavioral drift, with anomaly detection improving by 85.43% and behavior prediction by 70.51% on average. The code is available at https://github.com/horizonsinzqs/SmartGen.

cs.AI↗

arXiv2025

D2PPO: Diffusion Policy Policy Optimization with Dispersive Loss

Diffusion policies excel at robotic manipulation by naturally modeling multimodal action distributions in high-dimensional spaces. Nevertheless, diffusion policies suffer from diffusion representation collapse: semantically similar observations are mapped to indistinguishable features, ultimately impairing their ability to handle subtle but critical variations required for complex robotic manipulation. To address this problem, we propose D2PPO (Diffusion Policy Policy Optimization with Dispersive Loss). D2PPO introduces dispersive loss regularization that combats representation collapse by treating all hidden representations within each batch as negative pairs. D2PPO compels the network to learn discriminative representations of similar observations, thereby enabling the policy to identify subtle yet crucial differences necessary for precise manipulation. In evaluation, we find that early-layer regularization benefits simple tasks, while late-layer regularization sharply enhances performance on complex manipulation tasks. On RoboMimic benchmarks, D2PPO achieves an average improvement of 22.7% in pre-training and 26.1% after fine-tuning, setting new SOTA results. In comparison with SOTA, results of real-world experiments on a Franka Emika Panda robot show the excitingly high success rate of our method. The superiority of our method is especially evident in complex tasks. Project page: https://guowei-zou.github.io/d2ppo/

cs.AI↗

arXiv2025

End-to-end image compression and reconstruction with ultrahigh speed and ultralow energy enabled by opto-electronic computing processor

The rapid development of AR/VR, remote sensing, satellite radar, and medical equipment has created an imperative demand for ultra efficient image compression and reconstruction that exceed the capabilities of electronic processors. For the first time, we demonstrate an end to end image compression and reconstruction approach using an optoelectronic computing processor,achieving orders of magnitude higher speed and lower energy consumption than electronic counterparts. At its core is a 32X32 silicon photonic computing chip, which monolithically integrates 32 high speed modulators, 32 detectors, and a programmable photonic matrix core, copackaged with all necessary control electronics (TIA, ADC, DAC, FPGA etc.). Leveraging the photonic matrix core programmability, the processor generates trainable compressive matrices, enabling adjustable image compression ratios (from 2X to 256X) to meet diverse application needs. Deploying a custom lightweight photonic integrated circuit oriented network (LiPICO-Net) enables high quality reconstruction of compressed images. Our approach delivers an end to end latency of only 49.5ps/pixel while consuming only less than 10.6nJ/pixel-both metrics representing 2-3 orders of magnitude improvement compared with classical models running on state-of-the-art GPUs. We validate the system on a 130 million-pixel aerial imagery, enabling real time compression where electronic systems falter due to power and latency constraints. This work not only provides a transformative solution for massive image processing but also opens new avenues for photonic computing applications.

physics.optics↗

arXiv2025

InceptionMamba: An Efficient Hybrid Network with Large Band Convolution and Bottleneck Mamba

Within the family of convolutional neural networks, InceptionNeXt has shown excellent competitiveness in image classification and a number of downstream tasks. Built on parallel one-dimensional strip convolutions, however, it suffers from limited ability of capturing spatial dependencies along different dimensions and fails to fully explore spatial modeling in local neighborhood. Besides, inherent locality constraints of convolution operations are detrimental to effective global context modeling. To overcome these limitations, we propose a novel backbone architecture termed InceptionMamba in this study. More specifically, the traditional one-dimensional strip convolutions are replaced by orthogonal band convolutions in our InceptionMamba to achieve cohesive spatial modeling. Furthermore, global contextual modeling can be achieved via a bottleneck Mamba module, facilitating enhanced cross-channel information fusion and enlarged receptive field. Extensive evaluations on classification and various downstream tasks demonstrate that the proposed InceptionMamba achieves state-of-the-art performance with superior parameter and computational efficiency. The source code will be available at https://github.com/Wake1021/InceptionMamba.

cs.CV↗

arXiv2025

OpenLKA: An Open Dataset of Lane Keeping Assist from Recent Car Models under Real-world Driving Conditions

Lane Keeping Assist (LKA) is widely adopted in modern vehicles, yet its real-world performance remains underexplored due to proprietary systems and limited data access. This paper presents OpenLKA, the first open, large-scale dataset for LKA evaluation and improvement. It includes 400 hours of driving data from 62 production vehicle models, collected through extensive road testing in Tampa, Florida and global contributions from the Comma.ai driving community. The dataset spans a wide range of challenging scenarios, including complex road geometries, degraded lane markings, adverse weather, lighting conditions and surrounding traffic. The dataset is multimodal, comprising: i) full CAN bus streams, decoded using custom reverse-engineered DBC files to extract key LKA events (e.g., system disengagements, lane detection failures); ii) synchronized high-resolution dash-cam video; iii) real-time outputs from Openpilot, providing accurate estimates of road curvature and lane positioning; iv) enhanced scene annotations generated by Vision Language Models, describing lane visibility, pavement quality, weather, lighting, and traffic conditions. By integrating vehicle-internal signals with high-fidelity perception and rich semantic context, OpenLKA provides a comprehensive platform for benchmarking the real-world performance of production LKA systems, identifying safety-critical operational scenarios, and assessing the readiness of current road infrastructure for autonomous driving. The dataset is publicly available at: https://github.com/OpenLKA/OpenLKA.

cs.CV↗

arXiv2025

Towards Automated Self-Supervised Learning for Truly Unsupervised Graph Anomaly Detection

Self-supervised learning (SSL) is an emerging paradigm that exploits supervisory signals generated from the data itself, and many recent studies have leveraged SSL to conduct graph anomaly detection. However, we empirically found that three important factors can substantially impact detection performance across datasets: 1) the specific SSL strategy employed; 2) the tuning of the strategy's hyperparameters; and 3) the allocation of combination weights when using multiple strategies. Most SSL-based graph anomaly detection methods circumvent these issues by arbitrarily or selectively (i.e., guided by label information) choosing SSL strategies, hyperparameter settings, and combination weights. While an arbitrary choice may lead to subpar performance, using label information in an unsupervised setting is label information leakage and leads to severe overestimation of a method's performance. Leakage has been criticized as "one of the top ten data mining mistakes", yet many recent studies on SSL-based graph anomaly detection have been using label information to select hyperparameters. To mitigate this issue, we propose to use an internal evaluation strategy (with theoretical analysis) to select hyperparameters in SSL for unsupervised anomaly detection. We perform extensive experiments using 10 recent SSL-based graph anomaly detection algorithms on various benchmark datasets, demonstrating both the prior issues with hyperparameter selection and the effectiveness of our proposed strategy.

cs.LG↗

arXiv2025

SlideCoder: Layout-aware RAG-enhanced Hierarchical Slide Generation from Design

Manual slide creation is labor-intensive and requires expert prior knowledge. Existing natural language-based LLM generation methods struggle to capture the visual and structural nuances of slide designs. To address this, we formalize the Reference Image to Slide Generation task and propose Slide2Code, the first benchmark with difficulty-tiered samples based on a novel Slide Complexity Metric. We introduce SlideCoder, a layout-aware, retrieval-augmented framework for generating editable slides from reference images. SlideCoder integrates a Color Gradient-based Segmentation algorithm and a Hierarchical Retrieval-Augmented Generation method to decompose complex tasks and enhance code generation. We also release SlideMaster, a 7B open-source model fine-tuned with improved reverse-engineered data. Experiments show that SlideCoder outperforms state-of-the-art baselines by up to 40.5 points, demonstrating strong performance across layout fidelity, execution accuracy, and visual consistency. Our code is available at https://github.com/vinsontang1/SlideCoder.

cs.CV↗

arXiv2025

QueryAttack: Jailbreaking Aligned Large Language Models Using Structured Non-natural Query Language

Recent advances in large language models (LLMs) have demonstrated remarkable potential in the field of natural language processing. Unfortunately, LLMs face significant security and ethical risks. Although techniques such as safety alignment are developed for defense, prior researches reveal the possibility of bypassing such defenses through well-designed jailbreak attacks. In this paper, we propose QueryAttack, a novel framework to examine the generalizability of safety alignment. By treating LLMs as knowledge databases, we translate malicious queries in natural language into structured non-natural query language to bypass the safety alignment mechanisms of LLMs. We conduct extensive experiments on mainstream LLMs, and the results show that QueryAttack not only can achieve high attack success rates (ASRs), but also can jailbreak various defense methods. Furthermore, we tailor a defense method against QueryAttack, which can reduce ASR by up to $64\%$ on GPT-4-1106. Our code is available at https://github.com/horizonsinzqs/QueryAttack.

cs.CR↗

arXiv2025

StepSearch: Igniting LLMs Search Ability via Step-Wise Proximal Policy Optimization

Efficient multi-hop reasoning requires Large Language Models (LLMs) based agents to acquire high-value external knowledge iteratively. Previous work has explored reinforcement learning (RL) to train LLMs to perform search-based document retrieval, achieving notable improvements in QA performance, but underperform on complex, multi-hop QA resulting from the sparse rewards from global signal only. To address this gap in existing research, we introduce StepSearch, a framework for search LLMs that trained with step-wise proximal policy optimization method. It consists of richer and more detailed intermediate search rewards and token-level process supervision based on information gain and redundancy penalties to better guide each search step. We constructed a fine-grained question-answering dataset containing sub-question-level search trajectories based on open source datasets through a set of data pipeline method. On standard multi-hop QA benchmarks, it significantly outperforms global-reward baselines, achieving 11.2% and 4.2% absolute improvements for 3B and 7B models over various search with RL baselines using only 19k training data, demonstrating the effectiveness of fine-grained, stepwise supervision in optimizing deep search LLMs. Our code will be released on https://github.com/Zillwang/StepSearch.

cs.CL↗

arXiv2025

Thinking Short and Right Over Thinking Long: Serving LLM Reasoning Efficiently and Accurately

Recent advances in test-time scaling suggest that Large Language Models (LLMs) can gain better capabilities by generating Chain-of-Thought reasoning (analogous to human thinking) to respond a given request, and meanwhile exploring more reasoning branches (i.e., generating multiple responses and ensembling them) can improve the final output quality. However, when incorporating the two scaling dimensions, we find that the system efficiency is dampened significantly for two reasons. Firstly, the time cost to generate the final output increases substantially as many reasoning branches would be trapped in the over-thinking dilemma, producing excessively long responses. Secondly, generating multiple reasoning branches for each request increases memory consumption, which is unsuitable for LLM serving since we can only batch a limited number of requests to process simultaneously. To address this, we present SART, a serving framework for efficient and accurate LLM reasoning. The essential idea is to manage the thinking to be short and right, rather than long. For one thing, we devise a redundant sampling with early stopping approach based on empirical observations and theoretic analysis, which increases the likelihood of obtaining short-thinking responses when sampling reasoning branches. For another, we propose to dynamically prune low-quality branches so that only right-thinking branches are maintained, reducing the memory consumption and allowing us to batch more requests. Experimental results demonstrate that SART not only improves the accuracy of LLM reasoning but also enhances the serving efficiency, outperforming existing methods by up to 28.2 times and on average 15.7 times in terms of efficiency when achieving the same level of accuracy.

cs.LG↗