检索结果-内蒙古大学图书馆

Improving Bert Fine-Tuning via Stabilizing Cross-Layer Mutual Information

学校读者我要写书评

暂无评论

Improving Bert Fine-Tuning via Stabilizing Cross-Layer Mutua...

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Jicun Li Xingjian Li Tianyang Wang Shi Wang Yanan Cao Chengzhong Xu Dejing Dou Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences University of Chinese Academy of Sciences Big Data Lab Baidu Research State Key Lab of IOTSC University of Macau University of Alabama at Birmingham Institute of Information Engineering Chinese Academy of Sciences

Fine-tuning pre-trained language models, such as BERT, has shown enormous success among various NLP tasks. Though simple and effective, the process of fine-tuning has been found unstable, which often leads to unexpected poor performance. To increase stability and generalizability, most existing works resort to maintaining the parameters or representations of pre-trained models during fine-tuning. Nevertheless, very little work explores mining the reliable part of pre-learned information that can help to stabilize fine-tuning. To address this challenge, we introduce a novel solution in which we fine-tune BERT with stabilized cross-layer mutual information. Our method aims to preserve the reliable behaviors of cross-layer information propagation, instead of preserving the information itself, of the pre-trained model. Therefore, our method circumvents the domain conflicts between pre-trained and target tasks. We conduct extensive experiments with popular pre-trained BERT variants on NLP datasets, demonstrating the universal effectiveness and robustness of our method.

关键词： Cross layer design Bit error rate Signal processing Stability analysis Robustness Acoustics Behavioral sciences

SG-NeRF: Neural Surface Reconstruction with Scene Graph Optimization

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Chen, Yiyang Dong, Siyan Wang, Xulong Cai, Lulu Zheng, Youyi Yang, Yanchao State Key Lab of CAD&CG Zhejiang University China Chohotech Co. Ltd. Institute of Data Science The University of Hong Kong Hong Kong Department of Electrical and Electronic Engineering The University of Hong Kong Hong Kong

3D surface reconstruction from images is essential for numerous applications. Recently, Neural Radiance Fields (NeRFs) have emerged as a promising framework for 3D modeling. However, NeRFs require accurate camera poses as input, and existing methods struggle to handle significantly noisy pose estimates (i.e., outliers), which are commonly encountered in real-world scenarios. To tackle this challenge, we present a novel approach that optimizes radiance fields with scene graphs to mitigate the influence of outlier poses. Our method incorporates an adaptive inlier-outlier confidence estimation scheme based on scene graphs, emphasizing images of high compatibility with the neighborhood and consistency in the rendering quality. We also introduce an effective intersection-over-union (IoU) loss to optimize the camera pose and surface geometry, together with a coarse-to-fine strategy to facilitate the training. Furthermore, we propose a new dataset containing typical outlier poses for a detailed evaluation. Experimental results on various datasets consistently demonstrate the effectiveness and superiority of our method over existing approaches, showcasing its robustness in handling outliers and producing high-quality 3D reconstructions. Our code and data are available at: https://***/Iris-cyy/SG-NeRF. Copyright © 2024, The Authors. All rights reserved.

关键词： Cameras

Validity Matters: Uncertainty-Guided Testing of Deep Neural Networks

学校读者我要写书评

暂无评论

Validity Matters: Uncertainty-Guided Testing of Deep Neural ...

作者： Jiang, Zhouxian Li, Honghui Wang, Rui Tian, Xuetao Liang, Ci Yan, Fei Zhang, Junwen Liu, Zhen School of Computer Science and Technology Beijing Jiaotong University Beijing China Beijing Key Lab of Traffic Data Analysis and Mining Beijing Jiaotong University Beijing China Collaborative Innovation Center of Railway Traffic Safety Beijing Jiaotong University Beijing China Faculty of Psychology Beijing Normal University Beijing China School of Transportation Science and Engineering Harbin Institute of Technology Harbin China School of Automation and Intelligence Beijing Jiaotong University Beijing China

Despite numerous applications of deep learning technologies on critical tasks in various domains, advanced deep neural networks (DNNs) face persistent safety and security challenges, such as the overconfidence in predicting out-of-distribution samples and susceptibility to adversarial examples. Thorough testing by exploring the input space serves as a key strategy to ensure their robustness and trustworthiness of these networks. However, existing testing methods focus on disclosing more erroneous model behaviours, overlooking the validity of the generated test inputs. To mitigate this issue, we investigate devising valid test input generation method for DNNs from a predictive uncertainty perspective. Through a large-scale empirical study across 11 predictive uncertainty metrics for DNNs, we explore the correlation between validity and uncertainty of test inputs. Our findings reveal that the predictive entropy-based and ensemble-based uncertainty metrics effectively characterize the input validity demonstration. Building on these insights, we introduce UCTest, an uncertainty-guided deep learning testing approach, to efficiently generate valid and authentic test inputs. We formulate a joint optimization objective: to uncover the model's misbehaviours by maximizing the loss function and concurrently generate valid test input by minimizing uncertainty. Extensive experiments demonstrate that our approach outperforms the current testing methods in generating valid test inputs. Furthermore, incorporating natural variation through data augmentation techniques into UCTest effectively boosts the diversity of generated test inputs. © 2024 John Wiley & Sons Ltd.

关键词： Adversarial machine learning

Detecting Backdoors During the Inference Stage Based on Corruption Robustness Consistency

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Liu, Xiaogeng Li, Minghui Wang, Haoyu Hu, Shengshan Ye, Dengpan Jin, Hai Wu, Libing Xiao, Chaowei School of Cyber Science and Engineering Huazhong University of Science and Technology China School of Computer Science and Technology Huazhong University of Science and Technology China School of Software Engineering Huazhong University of Science and Technology China National Engineering Research Center for Big Data Technology and System China Services Computing Technology and System Lab China Hubei Key Laboratory of Distributed System Security China Hubei Engineering Research Center on Big Data Security China Cluster and Grid Computing Lab China School of Cyber Science and Engineering Wuhan University China Arizona State University United States

Deep neural networks are proven to be vulnerable to backdoor attacks. Detecting the trigger samples during the inference stage, i.e., the test-time trigger sample detection, can prevent the backdoor from being triggered. However, existing detection methods often require the defenders to have high accessibility to victim models, extra clean data, or knowledge about the appearance of backdoor triggers, limiting their practicality. In this paper, we propose the test-time corruption robustness consistency evaluation (TeCo)1, a novel test-time trigger sample detection method that only needs the hard-label outputs of the victim models without any extra information. Our journey begins with the intriguing observation that the backdoor-infected models have similar performance across different image corruptions for the clean images, but perform discrepantly for the trigger samples. Based on this phenomenon, we design TeCo to evaluate test-time robustness consistency by calculating the deviation of severity that leads to predictions' transition across different corruptions. Extensive experiments demonstrate that compared with state-of-the-art defenses, which even require either certain information about the trigger types or accessibility of clean data, TeCo outperforms them on different backdoor attacks, datasets, and model architectures, enjoying a higher AUROC by 10% and 5 times of stability. Copyright © 2023, The Authors. All rights reserved.

关键词： Crime

KADEL: Knowledge-Aware Denoising Learning for Commit Message Generation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Tao, Wei Zhou, Yucheng Wang, Yanlin Zhang, Hongyu Wang, Haofen Zhang, Wenqiang Shanghai Engineering Research Center of AI and Robotics Academy for Engineering and Technology Fudan University Shanghai China State Key Laboratory of Internet of Things for Smart City Department of Computer and Information Science University of Macau China School of Software Engineering Sun Yat-sen University Guangdong Zhuhai519082 China School of Big Data and Software Engineering Chongqing University Chongqing China College of Design and Innovation Tongji University Shanghai China Engineering Research Center of AI and Robotics Ministry of Education Academy for Engineering and Technology Shanghai Key Lab of Intelligent Information Processing School of Computer Science Fudan University 220 Handan Road Shanghai200433 China

Commit messages are natural language descriptions of code changes, which are important for software evolution such as code understanding and maintenance. However, previous methods are trained on the entire dataset without considering the fact that a portion of commit messages adhere to good practice (i.e., good-practice commits), while the rest do not. On the basis of our empirical study, we discover that training on good-practice commits significantly contributes to the commit message generation. Motivated by this finding, we propose a novel knowledge-aware denoising learning method called KADEL. Considering that good-practice commits constitute only a small proportion of the dataset, we align the remaining training samples with these good-practice commits. To achieve this, we propose a model that learns the commit knowledge by training on good-practice commits. This knowledge model enables supplementing more information for training samples that do not conform to good practice. However, since the supplementary information may contain noise or prediction errors, we propose a dynamic denoising training method. This method composes a distribution-aware confidence function and a dynamic distribution list, which enhances the effectiveness of the training process. Experimental results on the whole MCMD dataset demonstrate that our method overall achieves state-of-the-art performance compared with previous methods. © 2024, CC BY.

关键词： Sampling

Localization of Image Splicing Under Segment Anything Model With Integrated Compression and Edge Artifacts

学校读者我要写书评

暂无评论

Localization of Image Splicing Under Segment Anything Model ...

IEEE International Conference on Image Processing

作者： Ruhao Zhao Xian Zhong Liang Liao Wenxuan Liu Wenxin Huang Zheng Wang School of Computer Science and Artificial Intelligence Hubei Key Laboratory of Transportation Internet of Things Wuhan University of Technology School of Electrical and Electronic Engineering Rapid-Rich Object Search Lab Nanyang Technological University College of Computing and Data Science Nanyang Technological University School of Computer Science and Information Engineering Hubei University School of Computer Science National Engineering Research Center for Multimedia Software Wuhan University

ISBN: (数字)9798350349399

ISBN: (纸本)9798350349405

The localization of image splicing involves identifying pixels in an image that have been spliced from other images, necessitating the discernment of splicing features. Despite significant advancements driven by the rise of social media and deep learning, existing methods exhibit limitations, often neglecting the integration of coarse and precise features and lacking the ability to understand objects. This leads to erroneous predictions in identifying spliced regions. This paper proposes Segment Anything Model with Integrated Compression and Edge artifacts (SAM-ICE) for the localization of image splicing, addressing these limitations by fusing forged edge features and compression artifact features. Leveraging SAM’s object understanding ability, our method identifies spliced regions using the fused features as guidance. Specifically, we employ Edge Artifact Extractor (EAE) to extract fine high-frequency edge splicing features and Compression Artifact Extractor (CAE) to extract coarse compression artifact features. By combining these features, our method utilizes coarse-fine features to accurately pinpoint the spliced portions of the image. Experimental results demonstrate the superior accuracy, robustness, and generalizability of our method compared to the state-of-the-arts.

关键词： Location awareness Image segmentation Image coding Social networking (online) Image edge detection Splicing Feature extraction

A Multi-Step Speed Prediction Method for Traffic Networks Based on MAF-GCN-Seq2Seq Model 2

学校读者我要写书评

暂无评论

A Multi-Step Speed Prediction Method for Traffic Networks Ba...

2nd International Conference on Electronic Information Technology and Smart Agriculture, ICEITSA 2022

作者： Long, Yi Xiao, Hongbo Xiao, Jianhua Peng, Xiaoning School of Computer Science and Engineering Huaihua University Hunan Huaihua418000 China Intelligent Processing and Application in Hunan Province Universities Key Laboratory of Wuling-Mountain Health Big Data Hunan Huaihua418000 China Key Lab. of Intelligent Contr. Technol. for Wuling-Mountain Ecological Agriculture in Hunan Province Hunan Huaihua418000 China

ISBN: (纸本)9781665464017

Accurately predicting the future trend of traffic flow is of great significance to the planning of urban traffic. The great challenge of predicting traffic flow stems from the complexity and uncertainty of traffic flow. To address the complexity of traffic flow, this paper proposes a MAF-GCN 2 Seq model. The model uses a moving average filter to filter the time series data, extracts the spatial features between each node through a graph convolutional neural network, and forms a time series matrix for input into the encoder (GRU), and introduces a self-attention mechanism to mine the spatial features. time-domain features for multi-step traffic speed prediction. In order to verify the advanced nature of the model, this paper uses the traffic flow data set in Shenzhen to predict the traffic flow. Through the comparison experiments, it can be seen that the prediction accuracy of the MAF-GCN 2 Seq model reaches 90%, which has obvious advantages compared with other models. © 2022 IEEE.

关键词： Convolutional neural networks

Denial-of-Service or Fine-Grained Control: Towards Flexible Model Poisoning Attacks on Federated Learning

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Zhang, Hangtao Yao, Zeming Zhang, Leo Yu Hu, Shengshan Chen, Chao Liew, Alan Li, Zhetao School of Cyber Science and Engineering Huazhong University of Science and Technology China Swinburne University of Technology Australia Griffith University Australia National Engineering Research Center for Big Data Technology and System China Services Computing Technology and System Lab. Hubei Key Laboratory of Distributed System Security China Hubei Engineering Research Center on Big Data Security China RMIT University Australia Xiangtan University China

Federated learning (FL) is vulnerable to poisoning attacks, where adversaries corrupt the global aggregation results and cause denial-of-service (DoS). Unlike recent model poisoning attacks that optimize the amplitude of malicious perturbations along certain prescribed directions to cause DoS, we propose a Flexible Model Poisoning Attack (FMPA) that can achieve versatile attack goals. We consider a practical threat scenario where no extra knowledge about the FL system (e.g., aggregation rules or updates on benign devices) is available to adversaries. FMPA exploits the global historical information to construct an estimator that predicts the next round of the global model as a benign reference. It then fine-tunes the reference model to obtain the desired poisoned model with low accuracy and small perturbations. Besides the goal of causing DoS, FMPA can be naturally extended to launch a fine-grained controllable attack, making it possible to precisely reduce the global accuracy. Armed with precise control, malicious FL service providers can gain advantages over their competitors without getting noticed, hence opening a new attack surface in FL other than DoS. Even for the purpose of DoS, experiments show that FMPA significantly decreases the global accuracy, outperforming six state-of-the-art attacks. © 2023, CC BY.

关键词： Federated learning

Hybrid Parallel Inference for Large Model on Heterogeneous Clusters for High Throughput

学校读者我要写书评

暂无评论

Hybrid Parallel Inference for Large Model on Heterogeneous C...

International Conference on Parallel and Distributed Systems (ICPADS)

作者： Chaonong Xu Weiming Kong Min Liu Mingming Zhang Chao Li Luqi Gong Beijing Key Lab of Petroleum Data Mining China University of Petroleum (Beijing) Beijing China Institute of Intelligent Computing Zhejiang Lab Hangzhou China School of Computer Science (National Pilot Software Engineering School) BUPT Beijing China

In high-throughput intelligent computing scenarios, multi-device parallelism strategies based on data parallelism or pipeline parallelism have been extensively utilized to accelerate large deep neural network model inference. data parallelism offers nearly linear improvement in inference speed, but it is limited by the memory capacity of a single device which constrains the model size. On the other hand, pipeline parallelism can support larger models, but the total communication of the activations among devices is high, which limits the improvement of the inference speed. To address the demand for efficient model inference in high-throughput heterogeneous scenarios, we proposes a hybrid parallelism strategy that combines data parallelism and pipeline parallelism. The strategy involves grouping heterogeneous device clusters and then employing inter-group data parallelism along with intra-group pipeline parallelism. Moreover, we propose an algorithm to find an optimal hybrid parallel inference strategy with maximum throughput. The control variables of the strategy includes the number of groups, group-device assignments and model partition ratios. Our experimental evaluation demonstrates that compared to PipeEdge, a pipeline parallel inference framework for heterogeneous cluster, our strategy can achieve 1.7× −3.4× acceleration in an 8-device heterogeneous cluster without loss of accuracy.

关键词：