PyTorchKR Papers

“Code Generation”

총 10,163편
태그·기간 등 상세 필터

카테고리 탐색

arXiv:2303.06818공식 코드 · Papers with Code

Backdoor Defense via Deconfounded Representation Learning

Deep neural networks (DNNs) are recently shown to be vulnerable to backdoor attacks, where attackers embed hidden backdoors in the DNN model by injecting a few poisoned examples into the training dataset. While extensive efforts have been made to detect and remove backdoors from backdoored DNNs, it is still not clear whether a backdoor-free clean model can be directly obtained from poisoned datasets. In this paper, we first construct a causal graph to model the generation process of poisoned data and find that the backdoor attack acts as the confounder, which brings spurious associations between the input images and target labels, making the model predictions less reliable. Inspired by the causal understanding, we propose the Causality-inspired Backdoor Defense (CBD), to learn deconfounded representations for reliable classification. Specifically, a backdoored model is intentionally trained to capture the confounding effects. The other clean model dedicates to capturing the desired causal effects by minimizing the mutual information with the confounding representations from the backdoored model and employing a sample-wise re-weighting scheme. Extensive experiments on multiple benchmark datasets against 6 state-of-the-art attacks verify that our proposed defense method is effective in reducing backdoor threats while maintaining high accuracy in predicting benign samples. Further analysis shows that CBD can also resist potential adaptive attacks. The code is available at \url{https://github.com/zaixizhang/CBD}.

arXiv:2303.06681공식 코드 · Papers with Code

Learning Deep Intensity Field for Extremely Sparse-View CBCT Reconstruction

Sparse-view cone-beam CT (CBCT) reconstruction is an important direction to reduce radiation dose and benefit clinical applications. Previous voxel-based generation methods represent the CT as discrete voxels, resulting in high memory requirements and limited spatial resolution due to the use of 3D decoders. In this paper, we formulate the CT volume as a continuous intensity field and develop a novel DIF-Net to perform high-quality CBCT reconstruction from extremely sparse (fewer than 10) projection views at an ultrafast speed. The intensity field of a CT can be regarded as a continuous function of 3D spatial points. Therefore, the reconstruction can be reformulated as regressing the intensity value of an arbitrary 3D point from given sparse projections. Specifically, for a point, DIF-Net extracts its view-specific features from different 2D projection views. These features are subsequently aggregated by a fusion module for intensity estimation. Notably, thousands of points can be processed in parallel to improve efficiency during training and testing. In practice, we collect a knee CBCT dataset to train and evaluate DIF-Net. Extensive experiments show that our approach can reconstruct CBCT with high image quality and high spatial resolution from extremely sparse views within 1.6 seconds, significantly outperforming state-of-the-art methods. Our code will be available at https://github.com/xmed-lab/DIF-Net.

arXiv:2303.17393공식 코드 · Papers with Code

Dynamic Conceptional Contrastive Learning for Generalized Category Discovery

Generalized category discovery (GCD) is a recently proposed open-world problem, which aims to automatically cluster partially labeled data. The main challenge is that the unlabeled data contain instances that are not only from known categories of the labeled data but also from novel categories. This leads traditional novel category discovery (NCD) methods to be incapacitated for GCD, due to their assumption of unlabeled data are only from novel categories. One effective way for GCD is applying self-supervised learning to learn discriminate representation for unlabeled data. However, this manner largely ignores underlying relationships between instances of the same concepts (e.g., class, super-class, and sub-class), which results in inferior representation learning. In this paper, we propose a Dynamic Conceptional Contrastive Learning (DCCL) framework, which can effectively improve clustering accuracy by alternately estimating underlying visual conceptions and learning conceptional representation. In addition, we design a dynamic conception generation and update mechanism, which is able to ensure consistent conception learning and thus further facilitate the optimization of DCCL. Extensive experiments show that DCCL achieves new state-of-the-art performances on six generic and fine-grained visual recognition datasets, especially on fine-grained ones. For example, our method significantly surpasses the best competitor by 16.2% on the new classes for the CUB-200 dataset. Code is available at https://github.com/TPCD/DCCL.

arXiv:2305.12170공식 코드 · Papers with Code

Dual-Diffusion: Dual Conditional Denoising Diffusion Probabilistic Models for Blind Super-Resolution Reconstruction in RSIs

Previous super-resolution reconstruction (SR) works are always designed on the assumption that the degradation operation is fixed, such as bicubic downsampling. However, as for remote sensing images, some unexpected factors can cause the blurred visual performance, like weather factors, orbit altitude, etc. Blind SR methods are proposed to deal with various degradations. There are two main challenges of blind SR in RSIs: 1) the accu-rate estimation of degradation kernels; 2) the realistic image generation in the ill-posed problem. To rise to the challenge, we propose a novel blind SR framework based on dual conditional denoising diffusion probabilistic models (DDSR). In our work, we introduce conditional denoising diffusion probabilistic models (DDPM) from two aspects: kernel estimation progress and re-construction progress, named as the dual-diffusion. As for kernel estimation progress, conditioned on low-resolution (LR) images, a new DDPM-based kernel predictor is constructed by studying the invertible mapping between the kernel distribution and the latent distribution. As for reconstruction progress, regarding the predicted degradation kernels and LR images as conditional information, we construct a DDPM-based reconstructor to learning the mapping from the LR images to HR images. Com-prehensive experiments show the priority of our proposal com-pared with SOTA blind SR methods. Source Code is available at https://github.com/Lincoln20030413/DDSR

arXiv:2305.19101공식 코드 · Papers with Code

Which Models have Perceptually-Aligned Gradients? An Explanation via Off-Manifold Robustness

One of the remarkable properties of robust computer vision models is that their input-gradients are often aligned with human perception, referred to in the literature as perceptually-aligned gradients (PAGs). Despite only being trained for classification, PAGs cause robust models to have rudimentary generative capabilities, including image generation, denoising, and in-painting. However, the underlying mechanisms behind these phenomena remain unknown. In this work, we provide a first explanation of PAGs via \emph{off-manifold robustness}, which states that models must be more robust off- the data manifold than they are on-manifold. We first demonstrate theoretically that off-manifold robustness leads input gradients to lie approximately on the data manifold, explaining their perceptual alignment. We then show that Bayes optimal models satisfy off-manifold robustness, and confirm the same empirically for robust models trained via gradient norm regularization, randomized smoothing, and adversarial training with projected gradient descent. Quantifying the perceptual alignment of model gradients via their similarity with the gradients of generative models, we show that off-manifold robustness correlates well with perceptual alignment. Finally, based on the levels of on- and off-manifold robustness, we identify three different regimes of robustness that affect both perceptual alignment and model accuracy: weak robustness, bayes-aligned robustness, and excessive robustness. Code is available at \url{https://github.com/tml-tuebingen/pags}.

arXiv:2401.09417공식 코드 · Papers with Code

Vision Mamba: Efficient Visual Representation Learning with Bidirectional State Space Model

Recently the state space models (SSMs) with efficient hardware-aware designs, i.e., the Mamba deep learning model, have shown great potential for long sequence modeling. Meanwhile building efficient and generic vision backbones purely upon SSMs is an appealing direction. However, representing visual data is challenging for SSMs due to the position-sensitivity of visual data and the requirement of global context for visual understanding. In this paper, we show that the reliance on self-attention for visual representation learning is not necessary and propose a new generic vision backbone with bidirectional Mamba blocks (Vim), which marks the image sequences with position embeddings and compresses the visual representation with bidirectional state space models. On ImageNet classification, COCO object detection, and ADE20k semantic segmentation tasks, Vim achieves higher performance compared to well-established vision transformers like DeiT, while also demonstrating significantly improved computation & memory efficiency. For example, Vim is 2.8$\times$ faster than DeiT and saves 86.8% GPU memory when performing batch inference to extract features on images with a resolution of 1248$\times$1248. The results demonstrate that Vim is capable of overcoming the computation & memory constraints on performing Transformer-style understanding for high-resolution images and it has great potential to be the next-generation backbone for vision foundation models. Code is available at https://github.com/hustvl/Vim.

arXiv:2404.13039공식 코드 · Papers with Code

LaPA: Latent Prompt Assist Model For Medical Visual Question Answering

Medical visual question answering (Med-VQA) aims to automate the prediction of correct answers for medical images and questions, thereby assisting physicians in reducing repetitive tasks and alleviating their workload. Existing approaches primarily focus on pre-training models using additional and comprehensive datasets, followed by fine-tuning to enhance performance in downstream tasks. However, there is also significant value in exploring existing models to extract clinically relevant information. In this paper, we propose the Latent Prompt Assist model (LaPA) for medical visual question answering. Firstly, we design a latent prompt generation module to generate the latent prompt with the constraint of the target answer. Subsequently, we propose a multi-modal fusion block with latent prompt fusion module that utilizes the latent prompt to extract clinical-relevant information from uni-modal and multi-modal features. Additionally, we introduce a prior knowledge fusion module to integrate the relationship between diseases and organs with the clinical-relevant information. Finally, we combine the final integrated information with image-language cross-modal information to predict the final answers. Experimental results on three publicly available Med-VQA datasets demonstrate that LaPA outperforms the state-of-the-art model ARL, achieving improvements of 1.83%, 0.63%, and 1.80% on VQA-RAD, SLAKE, and VQA-2019, respectively. The code is publicly available at https://github.com/GaryGuTC/LaPA_model.

arXiv:2406.16072공식 코드 · Papers with Code

DV-3DLane: End-to-end Multi-modal 3D Lane Detection with Dual-view Representation

Accurate 3D lane estimation is crucial for ensuring safety in autonomous driving. However, prevailing monocular techniques suffer from depth loss and lighting variations, hampering accurate 3D lane detection. In contrast, LiDAR points offer geometric cues and enable precise localization. In this paper, we present DV-3DLane, a novel end-to-end Dual-View multi-modal 3D Lane detection framework that synergizes the strengths of both images and LiDAR points. We propose to learn multi-modal features in dual-view spaces, i.e., perspective view (PV) and bird's-eye-view (BEV), effectively leveraging the modal-specific information. To achieve this, we introduce three designs: 1) A bidirectional feature fusion strategy that integrates multi-modal features into each view space, exploiting their unique strengths. 2) A unified query generation approach that leverages lane-aware knowledge from both PV and BEV spaces to generate queries. 3) A 3D dual-view deformable attention mechanism, which aggregates discriminative features from both PV and BEV spaces into queries for accurate 3D lane detection. Extensive experiments on the public benchmark, OpenLane, demonstrate the efficacy and efficiency of DV-3DLane. It achieves state-of-the-art performance, with a remarkable 11.2 gain in F1 score and a substantial 53.5% reduction in errors. The code is available at \url{https://github.com/JMoonr/dv-3dlane}.

arXiv:2408.05846

A Universal Flexible Neuromorphic Tactile System with Multithreshold Strategy

Extremely increased unstructured data brought by the large-scale intelligent sensing devices application have big challenges not only in data storing and processing but also power consumption surging. Therefore, to improve energy efficiency and processing speed, a new generation system structure and construction strategy is necessary. Most biological nervous systems, especially the tactile system, have a good flexibility and data processing performance with low power usage. Inspired from this mechanism, to optimize the intelligent system, we report a universal fully flexible neuromorphic perception system with a strong compatibility and multi-threshold signal processing strategy by mimicking tactile nervous system. Peak signal accumulated from spike encoded sensor signal in front-end processing unit can be used for recognition task directly since the bionic synaptic plasticity. Compared with conventional systems, power consumption of our system significantly decreases about 1 order of magnitude in a same recognition task. What is more, the design of voltage-based matching circuit and multithreshold processing circuit provide an excellent compatibility and multi-signal processing capability in our system. In feasibility verification, our system can output trend of different input signals (continuous signal and frequency signal etc.) accurately and have a high recognition accuracy of 90% in the symbol pattern and 90% in Morse code. These properties of our neuromorphic system show a great application potential in intelligent devices and bionic robots.

arXiv:2410.15957공식 코드 · Papers with Code

CamI2V: Camera-Controlled Image-to-Video Diffusion Model

Recent advancements have integrated camera pose as a user-friendly and physics-informed condition in video diffusion models, enabling precise camera control. In this paper, we identify one of the key challenges as effectively modeling noisy cross-frame interactions to enhance geometry consistency and camera controllability. We innovatively associate the quality of a condition with its ability to reduce uncertainty and interpret noisy cross-frame features as a form of noisy condition. Recognizing that noisy conditions provide deterministic information while also introducing randomness and potential misguidance due to added noise, we propose applying epipolar attention to only aggregate features along corresponding epipolar lines, thereby accessing an optimal amount of noisy conditions. Additionally, we address scenarios where epipolar lines disappear, commonly caused by rapid camera movements, dynamic objects, or occlusions, ensuring robust performance in diverse environments. Furthermore, we develop a more robust and reproducible evaluation pipeline to address the inaccuracies and instabilities of existing camera control metrics. Our method achieves a 25.64% improvement in camera controllability on the RealEstate10K dataset without compromising dynamics or generation quality and demonstrates strong generalization to out-of-domain images. Training and inference require only 24GB and 12GB of memory, respectively, for 16-frame sequences at 256x256 resolution. We will release all checkpoints, along with training and evaluation code. Dynamic videos are best viewed at https://zgctroy.github.io/CamI2V.

arXiv:2502.08903

3D-Grounded Vision-Language Framework for Robotic Task Planning: Automated Prompt Synthesis and Supervised Reasoning

Vision-language models (VLMs) have achieved remarkable success in scene understanding and perception tasks, enabling robots to plan and execute actions adaptively in dynamic environments. However, most multimodal large language models lack robust 3D scene localization capabilities, limiting their effectiveness in fine-grained robotic operations. Additionally, challenges such as low recognition accuracy, inefficiency, poor transferability, and reliability hinder their use in precision tasks. To address these limitations, we propose a novel framework that integrates a 2D prompt synthesis module by mapping 2D images to point clouds, and incorporates a small language model (SLM) for supervising VLM outputs. The 2D prompt synthesis module enables VLMs, trained on 2D images and text, to autonomously extract precise 3D spatial information without manual intervention, significantly enhancing 3D scene understanding. Meanwhile, the SLM supervises VLM outputs, mitigating hallucinations and ensuring reliable, executable robotic control code generation. Our framework eliminates the need for retraining in new environments, thereby improving cost efficiency and operational robustness. Experimental results that the proposed framework achieved a 96.0\% Task Success Rate (TSR), outperforming other methods. Ablation studies demonstrated the critical role of both the 2D prompt synthesis module and the output supervision module (which, when removed, caused a 67\% TSR drop). These findings validate the framework's effectiveness in improving 3D recognition, task planning, and robotic task execution.

arXiv:2502.08658공식 코드 · Papers with Code

Knowledge-data fusion dominated vehicle platoon dynamics modeling and analysis: A physics-encoded deep learning approach

Recently, artificial intelligence (AI)-enabled nonlinear vehicle platoon dynamics modeling plays a crucial role in predicting and optimizing the interactions between vehicles. Existing efforts lack the extraction and capture of vehicle behavior interaction features at the platoon scale. More importantly, maintaining high modeling accuracy without losing physical analyzability remains to be solved. To this end, this paper proposes a novel physics-encoded deep learning network, named PeMTFLN, to model the nonlinear vehicle platoon dynamics. Specifically, an analyzable parameters encoded computational graph (APeCG) is designed to guide the platoon to respond to the driving behavior of the lead vehicle while ensuring local stability. Besides, a multi-scale trajectory feature learning network (MTFLN) is constructed to capture platoon following patterns and infer the physical parameters required for APeCG from trajectory data. The human-driven vehicle trajectory datasets (HIGHSIM) were used to train the proposed PeMTFLN. The trajectories prediction experiments show that PeMTFLN exhibits superior compared to the baseline models in terms of predictive accuracy in speed and gap. The stability analysis result shows that the physical parameters in APeCG is able to reproduce the platoon stability in real-world condition. In simulation experiments, PeMTFLN performs low inference error in platoon trajectories generation. Moreover, PeMTFLN also accurately reproduces ground-truth safety statistics. The code of proposed PeMTFLN is open source.

arXiv:2503.15264

LEGION: Learning to Ground and Explain for Synthetic Image Detection

The rapid advancements in generative technology have emerged as a double-edged sword. While offering powerful tools that enhance convenience, they also pose significant social concerns. As defenders, current synthetic image detection methods often lack artifact-level textual interpretability and are overly focused on image manipulation detection, and current datasets usually suffer from outdated generators and a lack of fine-grained annotations. In this paper, we introduce SynthScars, a high-quality and diverse dataset consisting of 12,236 fully synthetic images with human-expert annotations. It features 4 distinct image content types, 3 categories of artifacts, and fine-grained annotations covering pixel-level segmentation, detailed textual explanations, and artifact category labels. Furthermore, we propose LEGION (LEarning to Ground and explain for Synthetic Image detectiON), a multimodal large language model (MLLM)-based image forgery analysis framework that integrates artifact detection, segmentation, and explanation. Building upon this capability, we further explore LEGION as a controller, integrating it into image refinement pipelines to guide the generation of higher-quality and more realistic images. Extensive experiments show that LEGION outperforms existing methods across multiple benchmarks, particularly surpassing the second-best traditional expert on SynthScars by 3.31% in mIoU and 7.75% in F1 score. Moreover, the refined images generated under its guidance exhibit stronger alignment with human preferences. The code, model, and dataset will be released.

arXiv:2609.18128v2cs.AI

Symbolic Temporal Supervision of LLM Agents Using Contracts

Large language model (LLM) agents augmented by tools can automate complex, multi-step tasks, such as web navigation, code generation, and workflow orchestration, by acting on external systems through tool calls. However, hallucinations, distributional instability, and adversarial manipulations in LLMs, and the irreversible consequences of certain tool calls can lead to harmful outcomes. These risks call for safeguards that can check requirements across a trajectory, both before actions execute and when recorded behavior is evaluated. We present ContrAgent, a contract-based framework for symbolic temporal supervision of LLM agents. ContrAgent captures an agent's behavior as a sequence of tool calls and formalizes it as a trace over a fixed set of checkable predicates. It then specifies required behaviors using assume-guarantee contracts in linear temporal logic over finite traces (LTL$_f$). Each contract is compiled to a deterministic finite automaton (DFA) that serves two roles: gating agent actions online and evaluating recorded traces offline. A contract library, acting as a reusable knowledge base, is maintained independently of the agent's model and can be applied across different agents within the same task domain. We show the effectiveness of our approach on four benchmarks spanning both roles, where ContrAgent is competitive with LLM-judge, classifier, and guardrail baselines while producing deterministic, reproducible verdicts and, in the online mode, orders of magnitude lower per-call latency.

arXiv:2611.13309v1cs.DC

PegasusAI Studio: Legible Cyberinfrastructure and Accountable AI Agents for Scientific Workflows

AI coding agents can now construct scientific workflows, but operating them on shared cyberinfrastructure (CI) raises problems that code generation does not solve. Agents need current knowledge of the workflow system and its execution sites and access to the runtime state of distributed executions, while facilities need to attribute every agent action to a user and bound it by that user's privileges. We present PegasusAI Studio, an AI-assisted science gateway built on the Pegasus and HTCondor stack. Studio makes CI legible to agents by projecting one facility-curated knowledge base into the native configuration of several coding agents, by introspecting workflow interfaces and execution sites, and by grounding diagnosis in workflow and scheduler telemetry. Agent actions operate within a trust envelope that combines federated authentication, per-user execution environments and scheduler credentials, provider-neutral model access, observability, and approval controls for state-changing actions. The same application runs as a JupyterHub-based gateway on a research cloud and as an Open OnDemand application on HPC systems. We evaluate Studio through case studies of legacy CASA weather workflow modernization, a gravitational-wave reanalysis, an EEG analysis built with an open-weight model, and an Event Horizon Telescope (EHT) M87 image reconstruction that recovered from two failed runs to a zero-failure execution. The evaluation establishes that these mechanisms operate as designed.

arXiv:2611.19393v1cs.CL

사실에서 페르소나까지: 전문가 혼합을 이용한 대규모 언어 모델의 해석 가능한 역할 언러닝

From Facts to Personas: Interpretable Role Unlearning in LLMs via Mixture-of-Experts

대규모 언어 모델(LLM)은 여러 역할이 등장하는 대화와 페르소나 시뮬레이션에 널리 쓰인다. 그러나 역할마다 다른 언어적·행동적 양상을 잘 모방하는 능력은 개인정보 위험을 낳고 사회적 편향을 강화하거나, 안전하지 않거나 통제하기 어려운 출력을 만들 수 있다. 이를 해결하기 위해 특정 역할과 관련된 말투 및 지식을 선택적으로 잊게 하면서 나머지 역할의 언어 능력과 행동을 보존하는 역할극 언러닝 과제를 도입한다. 기존 언러닝 과제가 사실 정보의 제거에 집중한 데 비해, 이 과제는 페르소나 고유의 특성도 잊도록 요구해 모델의 출력을 더 세밀하게 제어한다. 본 연구는 역할들을 더 독립적인 전문가 부분공간에 매핑해 역할 분리와 구조적 해석 가능성을 개선하는 전문가 혼합 프레임워크 RoleMoE를 제안한다. 역할마다 고유한 특징을 더 잘 구분하기 위해 상호 보완적인 손실 두 가지를 적용한다. 특화 손실은 역할별로 서로 다른 전문가 분포를 사용하도록 유도하고, 요인 분리 손실은 전문가마다 서로 다른 역량 부분공간을 갖도록 한다. 이 분리 구조 덕분에 언러닝을 대상 역할의 전문가에 선택적으로 집중할 수 있다. 평가용 RoleBench-Unlearn도 제안한다. 여러 백본 LLM의 실험에서 RoleMoE는 역할을 정밀하고 통제 가능하게 잊게 하면서, 대상이 아닌 역할의 언어 품질, 지식, 페르소나 일관성을 더 잘 보존했다. 코드와 데이터셋은 https://github.com/mbzuai-nlp/ROLEMOE 에서 확인할 수 있다.

arXiv:2609.11188v1cs.CV

A Multi-View and Confusion-Guided Ensemble Framework for Robust Synthetic Image Attribution

Synthetic image attribution (SIA) has become increasingly important with the rapid advancement of text-to-image generation models. However, accurately identifying the source model of a generated image remains challenging due to the growing similarity among modern diffusion-based generators and the presence of diverse post-processing operations. In this report, we present a multi-view and confusion-guided ensemble framework for the Synthetic Image Attribution Challenge of the DLMMDD Workshop at ICANN 2026. Our approach integrates multiple complementary architectures, including FFT-ConvNeXt, DINOv2, CLIP, and Xception, to capture diverse attribution cues from frequency, semantic, and forensic perspectives. To improve robustness against unknown degradations and image manipulations, extensive data augmentation strategies are employed during training, simulating realistic post-processing operations such as compression, resizing, grayscale conversion, and blur. Furthermore, we analyze the confusion patterns of the ensemble model and observe severe ambiguity between Stable Diffusion 3 and Stable Diffusion 3.5. To address this issue, we introduce a dedicated binary expert classifier that is selectively activated under low-confidence conditions. We additionally apply class-adaptive confidence calibration to improve the discrimination of challenging classes such as Tencent Hunyuan. The proposed framework achieved 99.53% on the public leaderboard and 99.20% on the private leaderboard. The source code and implementation details are publicly available at https://github.com/ZOMIN28/SIA.

arXiv:2609.23889v3cs.CR

SyzHarness: LLM 합성 퍼징 하네스를 활용한 패치 기반 커널 버그 재현

SyzHarness: Patch-Based Kernel Bug Reproduction with LLM-Synthesized Fuzzing Harnesses

자동화된 커널(kernel) 취약점 재현은 버그 분류, 패치 검증(validation), 회귀 테스트에 필수적이지만, 여전히 효과적이고 효율적인 해결책이 부족합니다. 핵심 과제는 두 가지입니다. 재현기는 먼저 취약한 상태에 도달하는 데 필요한 트리거 스캐폴드(trigger scaffold)를 복구해야 하며, 실제로 버그를 유발하는 정확한 구체적 값들을 결정해야 합니다. 기존의 지향성 퍼징(directed fuzzing) 방식은 필요한 트리거 스캐폴드를 복구하는 데 비효과적이며, 대규모 언어 모델(LLM) 전용 생성은 구체적 값 탐색 및 런타임 비결정론을 처리하는 데 어려움을 겪어 불안정합니다. 본 논문에서는 패치 기반 Linux 커널 취약점 재현을 위해 LLM 추론과 커버리지 유도 퍼징(coverage-guided fuzzing)을 결합한 프레임워크인 SyzHarness를 설계합니다. 패치가 주어지면, SyzHarness는 코드 탐색 도구로 그라운딩(grounding)된 LLM 에이전트(agent)를 사용하여 필수 전제 설정 로직은 고정하고 Syzkaller가 변이할 수 있도록 불확실하고 버그에 중요한 입력 파라미터만 노출하는 매개변수화된 퍼징 하네스를 합성합니다. 그런 다음 SyzHarness는 이 하네스를 Syzkaller 호환 인터페이스로 변환하고 계층적 도달 가능성 피드백을 사용하여 반복적으로 개선합니다. 저자들은 트리거 가능한 실제 Linux 커널 취약점의 여러 데이터셋(dataset)에서 SyzHarness를 평가합니다. 100개의 KernelCTF 사례에서 SyzHarness는 78%의 버그 재현 성공률을 달성합니다. SyzDirect 벤치마크(benchmark)에서 SyzHarness는 73%의 버그 재현 성공률을 기록하여 이전의 지향성 그레이박스 퍼징보다 실질적으로 우수한 성능을 보입니다. 2026년 3월 이후 수정된 최근의 트리거 가능한 50개 syzbot 버그에 대해, SyzHarness는 수정 커밋만을 입력으로 사용하여 40/50(80%)을 재현합니다.

arXiv:2610.13470v1cs.CV

GraftOOD: 비디오 보상 최적화를 위한 이점 가중 정류에 분포 외 예시 접목

GraftOOD: Grafting Out-of-Distribution Exemplars into Advantage-Weighted Rectification for Video Reward Optimization

비디오 보상(reward) 최적화는 대리 보상을 증가시키는 동시에 비디오의 역동성을 감소시킬 수 있으며, 본 논문에서는 비디오 수준의 모션 품질 보상 하에서도 발생하는 이러한 불일치를 모션 붕괴(motion collapse)로 규명합니다. 이에 따라 선택된 프롬프트(prompt) 그룹에서 온라인 롤아웃의 일부를 분포 외(out-of-distribution) 비디오 예시로 대체하는 GraftOOD를 제안합니다. 공유 속도 회귀 목적함수는 그룹 상대적 보상 정류와 시퀀스 수준 예시 감독을 결합하여, 예시 생성 궤적이나 추가적인 지도 손실(loss) 계수를 필요로 하지 않습니다. 또한 점별 최적 필드를 두 개의 소스별 타깃 필드가 소스 사후 확률로 가중 결합된 형태로 특성화합니다. 외부 감독을 제공하기 위해 2,397개 프롬프트에 대해 형식이 일치하는 9,588개의 예시를 포함하는 GraftSet-Video를 구축했습니다. Wan2.1-1.3B에서 GraftOOD는 보상 전용 FlowAWR 대비 Dynamic Degree를 4.86에서 44.41로 끌어올렸으며, HPSv3는 8.10에서 7.32로 감소했고 VisionReward는 유사한 수준을 유지했습니다. 주요 설정에서 GraftOOD는 세 가지 예시 증강 방법 중 10개 지표 중 9개에서 우위를 보였습니다. 예시 빈도 및 그룹 내 개수에 대한 소거 연구를 통해, 예시 사용량 증가에 따른 일률적 이득보다는 역동성과 보상 간의 트레이드오프가 존재함을 확인했습니다. 이러한 결과는 측정된 더 높은 역동성이 생성된 비디오의 일관되고 시각적으로 타당한 모션과 동일시되지 않는다는 점을 유의하면서도, 외부 예시 감독이 스칼라 보상 피드백을 보완하는 역할을 뒷받침합니다. 코드는 https://github.com/Fuzm7/FlowAWR 에서 확인할 수 있습니다.

arXiv:2610.18342v1cs.CV

CityCode: Building Agent-Operable 3D Cities from Sparse Aerial Observations

Urban spatial intelligence calls for 3D cities that agents can measure, interpret, and modify across successive tasks. Maintaining such cities as persistent, code-based states requires lightweight acquisition and support for consistent state updates without repeated dense reconstruction. We therefore formulate a new setting for grounded urban world construction: generating large-scale, instance-level parameterized 3D cities from one low-altitude aerial RGB image per region and known camera parameters, while recovering metrically scaled buildings and explicit road geometry and connectivity. We introduce CityCode, an agent-operable representation that unifies metric geometry, urban topology, visual semantics, and 3D assets through persistent entity identities, with dependency-aware versioned execution for validated updates and historical restoration. We further develop Air2Code to construct this representation by combining visual structure extraction with adapted metric depth estimation. To address occlusion and incomplete coverage, recovered structure and observed appearance guide generative completion and subsequent structure-aware 3D generation for buildings and backgrounds. Experiments on CitySample-Aerial and UrbanScene3D demonstrate substantially more accurate metric structure recovery and superior visual fidelity to the target cities compared with the evaluated baselines. Across 50 task chains of 50 steps each, CityCode supports continued querying, editing, planning, and historical restoration over evolving city states. Together, CityCode and Air2Code turn sparse aerial observations into persistent, agent-operable urban worlds for long-horizon urban spatial intelligence.