检索结果-内蒙古大学图书馆

CodePurify: Defend Backdoor Attacks on Neural Code Models via Entropy-based Purification

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Mu, Fangwen Wang, Junjie Yu, Zhuohao Shi, Lin Wang, Song Li, Mingyang Wang, Qing State Key Laboratory of Intelligent Game Institute of Software Chinese Academy of Sciences Beijing China University of Chinese Academy of Sciences Beijing China Beihang University Beijing China Lassonde School of Engineering York University Toronto Canada

Neural code models have found widespread success in tasks pertaining to code intelligence, yet they are vulnerable to backdoor attacks, where an adversary can manipulate the victim model’s behavior by inserting triggers into the source code. Recent studies indicate that advanced backdoor attacks can achieve nearly 100% attack success rates on many software engineering tasks. However, effective defense techniques against such attacks remain insufficiently explored. In this study, we propose CODEPURIFY, a novel defense against backdoor attacks on code models through entropy-based purification. Entropy-based purification involves the process of precisely detecting and eliminating the possible triggers in the source code while preserving its semantic information. Within this process, CODEPURIFY first develops a confidence-driven entropy-based measurement to determine whether a code snippet is poisoned and, if so, locates the triggers. Subsequently, it purifies the code by substituting the triggers with benign tokens using a masked language model. We extensively evaluate CODEPURIFY against four advanced backdoor attacks across three representative tasks and two popular code models. The results show that CODEPURIFY significantly outperforms four commonly used defense baselines, improving average defense performance by at least 40%, 40%, and 12% across the three tasks, respectively. These findings highlight the potential of CODEPURIFY to serve as a robust defense against backdoor attacks on neural code models. Copyright © 2024, The Authors. All rights reserved.

关键词： Semantics

Sharing, Teaching and Aligning: Knowledgeable Transfer Learning for Cross-Lingual Machine Reading Comprehension

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Cao, Tingfeng Wang, Chengyu Tan, Chuanqi Huang, Jun Zhu, Jinhui School of Software Engineering South China University of Technology China Alibaba Group China Key Laboratory of Big Data and Intelligent Robot South China University of Technology Ministry of Education China

In cross-lingual language understanding, machine translation is often utilized to enhance the transferability of models across languages, either by translating the training data from the source language to the target, or from the target to the source to aid inference. However, in cross-lingual machine reading comprehension (MRC), it is difficult to perform a deep level of assistance to enhance cross-lingual transfer because of the variation of answer span positions in different languages. In this paper, we propose X-STA, a new approach for cross-lingual MRC. Specifically, we leverage an attentive teacher to subtly transfer the answer spans of the source language to the answer output space of the target. A Gradient-Disentangled Knowledge Sharing technique is proposed as an improved cross-attention block. In addition, we force the model to learn semantic alignments from multiple granularities and calibrate the model outputs with teacher guidance to enhance cross-lingual transferability. Experiments on three multi-lingual MRC datasets show the effectiveness of our method, outperforming state-of-the-art approaches. 1 Copyright © 2023, The Authors. All rights reserved.

关键词： Semantics

Shift from Texture-bias to Shape-bias: Edge Deformation-based Augmentation for Robust Object Recognition

学校读者我要写书评

暂无评论

Shift from Texture-bias to Shape-bias: Edge Deformation-base...

International Conference on Computer Vision (ICCV)

作者： Xilin He Qinliang Lin Cheng Luo Weicheng Xie Siyang Song Feng Liu Linlin Shen Computer Vision Institute School of Computer Science & Software Engineering Shenzhen University Shenzhen Institue of Artificial Intelligence & Robotics for Society Guangdong Key Laboratory of Intelligent Information Processing University of Leicester

Recent studies have shown the vulnerability of CNNs under perturbation noises, which is partially caused by the reason that the well-trained CNNs are too biased toward the object texture, i.e., they make predictions mainly based on texture cues. To reduce this texture-bias, current studies resort to learning augmented samples with heavily perturbed texture to make networks be more biased toward relatively stable shape cues. However, such methods usually fail to achieve real shape-biased networks due to the insufficient diversity of the shape cues. In this paper, we propose to augment the training dataset by generating semantically meaningful shapes and samples, via a shape deformation-based online augmentation, namely as SDbOA. The samples generated by our SDbOA have two main merits. First, the augmented samples with more diverse shape variations enable networks to learn the shape cues more elaborately, which encourages the network to be shape-biased. Second, semantic-meaningful shape-augmentation samples could be produced by jointly regularizing the generator with object texture and edge-guidance soft constraint, where the edges are represented more robustly with a self information guided map to better against the noises on them. Extensive experiments under various perturbation noises demonstrate the obvious superiority of our shape-bias-motivated model over the state of the arts in terms of robustness performance. Code is available at https://***/C0notSilly/-ICCV-23-Edge-Deformation-based-Online-Augmentation.

关键词：

Large Language Models (LLMs) Inference Offloading and Resource Allocation in Cloud-Edge Networks: An Active Inference Approach

学校读者我要写书评

暂无评论

Large Language Models (LLMs) Inference Offloading and Resour...

IEEE Conference on Vehicular Technology (VTC)

作者： Jingcheng Fang Ying He F. Richard Yu Jianqiang Li Victor C. Leung Shenzhen Key Lab of Digital & Intelligent Technologies & Systems College of Computer Science and Software Engineering Shenzhen University China Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ) Shenzhen China

As the research and applications of large language model (LLM) become increasingly sophisticated, it is difficult for resource-limited mobile terminals to run large-model inference tasks efficiently. Traditional deep reinforcement learning (DRL) based approaches have been used to offload LLM inference tasks to servers. However, existing solutions suffer from data inefficiency, insensitivity to latency requirements, and non-adaptability to task load variations. In this paper, we propose an active inference with rewardless guidance algorithm using expected future free energy for offloading decisions and allocating resources for the LLM inference task offloading and resource allocation problem of cloud-edge networks systems. Experimental results show that our proposed method has superior performance over mainstream DRLs, improves in data utilization efficiency, and is more adaptable to changing task load scenarios.

关键词：

DEGSTalk: Decomposed Per-Embedding Gaussian Fields for Hair-Preserving Talking Face Synthesis

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Deng, Kaijun Zheng, Dezhi Xie, Jindong Wang, Jinbao Xie, Weicheng Shen, Linlin Song, Siyang Computer Vision Institute School of Computer Science and Software Engineering Shenzhen University China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University China Guangdong Provincial Key Laboratory of Intelligent Information Processing China Department of Computer Science University of Exeter United Kingdom

Accurately synthesizing talking face videos and capturing fine facial features for individuals with long hair presents a significant challenge. To tackle these challenges in existing methods, we propose a decomposed per-embedding Gaussian fields (DEGSTalk), a 3D Gaussian Splatting (3DGS)-based talking face synthesis method for generating realistic talking faces with long hairs. Our DEGSTalk employs Deformable Pre-Embedding Gaussian Fields, which dynamically adjust pre-embedding Gaussian primitives using implicit expression coefficients. This enables precise capture of dynamic facial regions and subtle expressions. Additionally, we propose a Dynamic Hair-Preserving Portrait Rendering technique to enhance the realism of long hair motions in the synthesized videos. Results show that DEGSTalk achieves improved realism and synthesis quality compared to existing approaches, particularly in handling complex facial dynamics and hair preservation. Our code will be publicly available at https://***/CVI-SZU/DEGSTalk. Copyright © 2024, The Authors. All rights reserved.

关键词： Gaussian distribution

HairDiffusion: Vivid Multi-Colored Hair Editing via Latent Diffusion

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zeng, Yu Zhang, Yang Liu, Jiachen Shen, Linlin Deng, Kaijun He, Weizhao Wang, Jinbao Computer Vision Institute School of Computer Science & Software Engineering Shenzhen University China Shenzhen Institute of Artificial Intelligence and Robotics for Society China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University China Guangdong Provincial Key Laboratory of Intelligent Information Processing China

Hair editing is a critical image synthesis task that aims to edit hair color and hairstyle using text descriptions or reference images, while preserving irrelevant attributes (e.g., identity, background, cloth). Many existing methods are based on StyleGAN to address this task. However, due to the limited spatial distribution of StyleGAN, it struggles with multiple hair color editing and facial preservation. Considering the advancements in diffusion models, we utilize Latent Diffusion Models (LDMs) for hairstyle editing. Our approach introduces Multi-stage Hairstyle Blend (MHB), effectively separating control of hair color and hairstyle in diffusion latent space. Additionally, we train a warping module to align the hair color with the target region. To further enhance multi-color hairstyle editing, we fine-tuned a CLIP model using a multi-color hairstyle dataset. Our method not only tackles the complexity of multi-color hairstyles but also addresses the challenge of preserving original colors during diffusion editing. Extensive experiments showcase the superiority of our method in editing multi-color hairstyles while preserving facial attributes given textual descriptions and reference images. © 2024, CC BY.

关键词： Color image processing

Dynamic Spatio-Temporal Adaptive Convolutional Neural Networks for Traffic Flow Prediction

学校读者我要写书评

暂无评论

Dynamic Spatio-Temporal Adaptive Convolutional Neural Networ...

International Conference on intelligent Transportation

作者： Zhanbo Sun Zhuo Liu Chao Zhang Mingming Kong Ang Ji School of Transportation and Logistics Southwest Jiaotong University Chengdu Sichuan China Intelligent Policing Key Laboratory of Sichuan Province Sichuan Police College Luzhou Sichuan China School of Computer and Software Engineering Xihua University Chengdu China

ISBN: (数字)9798331505929

ISBN: (纸本)9798331505936

Traffic flow prediction is crucial for enhancing transportation systems' efficiency and safety, providing references for intelligent traffic management and road resource allocation. However, capturing the intricate spatial and temporal relationships between various areas and time intervals poses a significant challenge. To address the sophisticated spatio-temporal correlations encountered in traffic flow forecasting, we develop an adaptive convolutional neural network based on time attention. For spatial considerations, we design deformable convolutional kernels integrated with an attention mechanism to extract road information and adapt it to the road network structure. Furthermore, we deploy the attention mechanism to dynamically aggregate multiple parallel convolutional kernels instead of the traditional single-layer convolution on each input, focusing precisely on key areas within the network. From a temporal perspective, a dynamic time attention mechanism is applied based on weekly, daily, and nearest temporal data to grasp critical moments in periodic traffic flow. By integrating essential temporal and spatial elements, the Dynamic Spatial Temporal Adaptive Convolutional Networks (DSTACN) are constructed. The proposed model is validated on three real-world datasets. The results demonstrate its forecasting accuracy surpassing other baseline models by up to 11.98%, highlighting the proposed approach's capability in handling the spatio-temporal dependencies inherent in traffic flow data.

关键词： Adaptation models Attention mechanisms Adaptive systems Roads Transportation Safety Convolutional neural networks Resource management Kernel Forecasting

Quantum Reinforcement Learning with Quantum World Model

学校读者我要写书评

暂无评论

Quantum Reinforcement Learning with Quantum World Model

IEEE Conference on Global Communications (GLOBECOM)

作者： Peigen Zeng Ying He F. Richard Yu Victor C.M. Leung Shenzhen Key Lab of Digital & Intelligent Technologies & Systems College of Computer Science and Software Engineering Shenzhen University China Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ) Shenzhen China

Quantum reinforcement learning (QRL) can outperform classical reinforcement learning (RL) by utilizing quantum parallel theory and quantum phenomena such as superposition and entanglement. Although some excellent work has been done on QRL, most existing works either fail to show the exponential advantage of quantum computation over classical computation in terms of performance or are too demanding on quantum devices. In this paper, we provide a novel perspective on combining quantum computing and RL with faster convergence speed and relatively relaxed demands on quantum devices. Specifically, we propose a method to construct a world model with quantum circuit that allows it to interact in a quantum way. In addition, we use Grover's algorithm to efficiently extract high-value information from the quantum world model. Extensive simulation results show that the proposed method can have superior performance compared to classical RL algorithms.

关键词：

CA-Edit: Causality-Aware Condition Adapter for High-Fidelity Local Facial Attribute Editing

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Xian, Xiaole He, Xilin Niu, Zenghao Zhang, Junliang Xie, Weicheng Song, Siyang Yu, Zitong Shen, Linlin Computer Vision Institute School of Computer Science & Software Engineering Shenzhen University China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University China Guangdong Key Laboratory of Intelligent Information Processing China University of Exeter United Kingdom Great Bay University China

For efficient and high-fidelity local facial attribute editing, most existing editing methods either require additional finetuning for different editing effects or tend to affect beyond the editing regions. Alternatively, inpainting methods can edit the target image region while preserving external areas. However, current inpainting methods still suffer from the generation misalignment with facial attributes description and the loss of facial skin details. To address these challenges, (i) a novel data utilization strategy is introduced to construct datasets consisting of attribute-text-image triples from a data-driven perspective, (ii) a Causality-Aware Condition Adapter is proposed to enhance the contextual causality modeling of specific details, which encodes the skin details from the original image while preventing conflicts between these cues and textual conditions. In addition, a Skin Transition Frequency Guidance technique is introduced for the local modeling of contextual causality via sampling guidance driven by low-frequency alignment. Extensive quantitative and qualitative experiments demonstrate the effectiveness of our method in boosting both fidelity and editability for localized attribute editing. The code is available at https://***/connorxian/CA-Edit. © 2024, CC BY.

关键词：