Arrow Research search

Author name cluster

Yi Xie

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

19 papers
2 author rows

Possible papers

19

JBHI Journal 2026 Journal Article

Myocardial Infarction Detection with Incomplete Multi-View Data via Dual-Branch Gating Completion and Dirichlet Weighting

  • Yadi Wang
  • Yulin Xie
  • Yi Xie
  • Lin Chen
  • Bingbing Jiang

Multi-view myocardial infarction (MI) detection often relies on electrocardiogram (ECG) data, which suffers from noise sensitivity and limited early diagnostic value. Echocardiography provides richer temporal-spatial information but frequently encounters missing views due to clinical acquisition constraints. Existing fusion methods typically adopt static or overly complex dynamic weighting, limiting their adaptability to varying view quality and hindering real-time applicability. To this end, this paper proposes a view completion method based on a dual-branch gating structure, combining Transformer and Graph Neural Network (GNN) to complete incomplete multi-view information. Specifically, the model uses the Transformer encoder to model global temporal dependencies, introduces GNN to strengthen the local structural relationship between views, and uses the gating mechanism to achieve collaborative completion of the aforementioned two core components. Furthermore, this paper designs an uncertainty-driven dynamic weighted fusion strategy based on Dirichlet distribution, which can adaptively adjust the fusion weights according to the prediction confidence of each view, overcoming the limitations of traditional static weighting. Experiments on the HMC-QU dataset show that the proposed method achieves 92. 31% accuracy, 90. 00% precision, and 100. 00% specificity, outperforming state-of-the-art models and demonstrating strong potential for clinical deployment.

Details DOI

AAMAS Conference 2026 Conference Paper

SAT: Sequential Agent Tuning for Coordinator-Free Plug-and-Play Multi-LLM Training with Monotonic Improvement Guarantees

  • Yi Xie
  • Yangyang Xu
  • Yi Fan
  • Bo Liu

Large language models (LLMs) with a large number of parameters achieve strong performance but are often prohibitively expensive to deploy. Recent work explores using teams of smaller, more efficient LLMs that collectively match or even outperform a single large model. However, jointly updating multiple agents introduces compounding distribution shifts, making coordination and stability during training difficult. We address this by introducing Sequential Agent Tuning (SAT), a coordinator-free training paradigm. SAT represents the team as a factorized policy and employs blockcoordinate updates over agents, enabling scalable, decentralized training without a central controller. Specifically, we develop a sequence-aware, on-policy advantage estimator that conditions on the evolving team policy, coupled with per-agent KL trust regions that isolate occupancy drift. Theoretically, this framework provides two critical guarantees. First, it ensures monotonic improvement, stabilizing the training process. Second, it establishes provable plugand-playinvariance: anyagentcanbeupgradedtoastrongermodel without retraining the rest of the team, with a formal guarantee that the performance bound improves. Empirically, a team of three 4B agents (12B total) trained with SAT surpasses the much larger Qwen3-32B on AIME24/25 benchmarks by 3. 9% on average. We validate our plug-and-play theory by swapping in two 8B agents, which boosts the composite score by 10. 4%. We provide code and appendix of proof at https: //github. com/Yydc/SAT-AAMAS

PDF

ICML Conference 2025 Conference Paper

A Multi-Region Brain Model to Elucidate the Role of Hippocampus in Spatially Embedded Decision-Making

  • Yi Xie
  • Jaedong Hwang
  • Carlos D. Brody
  • David W. Tank
  • Ila Rani Fiete

Brains excel at robust decision-making and data-efficient learning. Understanding the architectures and dynamics underlying these capabilities can inform inductive biases for deep learning. We present a multi-region brain model that explores the normative role of structured memory circuits in a spatially embedded binary decision-making task from neuroscience. We counterfactually compare the learning performance and neural representations of reinforcement learning (RL) agents with brain models of different interaction architectures between grid and place cells in the entorhinal cortex and hippocampus, coupled with an action-selection cortical recurrent neural network. We demonstrate that a specific architecture–where grid cells receive and jointly encode self-movement velocity signals and decision evidence increments–optimizes learning efficiency while best reproducing experimental observations relative to alternative architectures. Our findings thus suggest brain-inspired structured architectures for efficient RL. Importantly, the models make novel, testable predictions about organization and information flow within the entorhinal-hippocampal-neocortical circuit: we predict that grid cells must conjunctively encode position and evidence for effective spatial decision-making, directly motivating new neurophysiological experiments.

Details

AAMAS Conference 2025 Conference Paper

ACORN: Acyclic Coordination with Reachability Network to Reduce Communication Redundancy in Multi-Agent Systems

  • Yi Xie
  • Ziqing Zhou
  • Chun Ouyang
  • Siao Liu
  • Linqiang Hu
  • Zhongxue Gan

Effective communication is essential in multi-agent reinforcement learning (MARL) for coordinating actions and maximizing collective rewards. Two common approaches for establishing communication are Graph Neural Networks (GNNs) and Transformers. Both methods introduce communication redundancy in complex scenarios. GNN-based methods model agent relationships through entire graph structures, leading to increased computational time. Transformers also increase computations due to self-attention calculations at each node. In this study, the ACORN (Acyclic Coordination with Reachability Networks) framework was introduced, utilizing acyclic coordination combined with a reachability-based attention mechanism. The most relevant nodes and connections in the GNN graph are used for self-attention calculations. Time complexity is reduced to 𝑂(|𝑉 | × 𝑛𝑘 × 𝑑), which is significantly better than the 𝑂(|𝑉 |2𝑑) complexity of standard Transformers. Acyclicity is ensured through Auto-Regressive Policy Learning and Sequence- Based Critic Learning. Experiments demonstrate that ACORN outperforms state-of-the-art methods, achieving an average improvement of 11% over MAT in challenging SMACV2 tasks and a 17% improvement within the same training time and steps.

PDF

EAAI Journal 2025 Journal Article

Attention based network for real-time road drivable area, lane line detection and scene identification

  • Feng You
  • Yi Xie
  • Siyi Zhang
  • Hao Chen
  • Haiwei Wang
  • Wei Zhang
  • Jianrong Liu

The detection of road drivable areas and lane lines is considered a fundamental component of autonomous driving systems. However, most existing approaches handle these tasks independently, and multi-task networks frequently neglect the inherent correlation between them while failing to differentiate various lane line types. In practice, the delineation of drivable regions is strongly influenced by both lane line characteristics and contextual street scenes. To address these limitations, a novel multi-task network—Real-time Road Drivable Area, Lane Line Detection, and Scene Identification Network (RLSNet)—is proposed. This network is designed to perform simultaneous segmentation of drivable areas, detection of lane lines, and classification of road scenes. Drivable area estimation is optimized through the integration of lane and scene cues, guided by traffic regulations. A Residual Network (ResNet)-based backbone is employed, enhanced with Bidirectional Fusion Attention (BFA) for feature encoding. This is followed by a decoder incorporating a Feature Aggregation Module (FAM) to enable effective semantic–spatial fusion. Lane line detection is further refined using a Bilateral Up-Sampling Decoder (BUSD), while scene understanding is enhanced via a Scene Classification Module (SCM). Extensive experiments conducted on the challenging Berkeley DeepDrive 100K(BDD100K) dataset have demonstrated that RLSNet achieves high accuracy in both drivable area and lane line detection by leveraging the mutual guidance of lane and scene information. Furthermore, the network maintains real-time inference speed at 93 frames per second (FPS), striking a practical balance between semantic fidelity and computational efficiency for real-world deployment. The implementation code has been made publicly available at: https: //github. com/033186ZSY/RLSNet-master.

Details DOI

AAAI Conference 2025 Conference Paper

Enhancing Masked Time-Series Modeling via Dropping Patches

  • Tianyu Qiu
  • Yi Xie
  • Hao Niu
  • Yun Xiong
  • Xiaofeng Gao

This paper explores how to enhance existing masked time-series modeling by randomly dropping sub-sequence level patches of time series. On this basis, a simple yet effective method named DropPatch is proposed, which has two remarkable advantages: 1) It improves the pre-training efficiency by a square-level advantage; 2) It provides additional advantages for modeling in scenarios such as in-domain, cross-domain, few-shot learning and cold start. This paper conducts comprehensive experiments to verify the effectiveness of the method and analyze its internal mechanism. Empirically, DropPatch strengthens the attention mechanism, reduces information redundancy and serves as an efficient means of data augmentation. Theoretically, it is proved that DropPatch slows down the rate at which the Transformer representations collapse into the rank-1 linear subspace by randomly dropping patches, thus optimizing the quality of the learned representations.

PDF Details DOI

AAMAS Conference 2025 Conference Paper

Heuristics-Assisted Experience Replay Strategy for Cooperative Multi-Agent Reinforcement Learning

  • Yi Xie
  • Ziqing Zhou
  • Chun Ouyang
  • Siao Liu
  • Linqiang Hu
  • Zhongxue Gan

Cooperative Multi-agent Reinforcement Learning (CMARL) has great potential for developing coordinated strategies that optimize team performance. However, common methods often fail to properly separate and utilize individual experiences due to a lack of effective team reward decomposition. The Heuristics-assisted Experience Replay Strategy (HAER) addresses this by decomposing team rewards into individual rewards and enabling efficient experience replay in MARL. By maintaining network gradient invariance, we derive a partial differential equation for the individual reward function, allowing accurate calculation of TD-errors and experience importance. The Cooperative Multi-Objective Swarm Optimization (CMOSO) algorithm is used to balance TD-errors and individual rewards for efficient learning. Extensive experiments on benchmarks demonstrate HAER’s effectiveness, with up to a 17. 6% performance boost in the homogeneous SMACV2 scenario and an average 8% improvement in GRF for heterogeneous agent cooperation.

PDF

NeurIPS Conference 2025 Conference Paper

MemSim: A Bayesian Simulator for Evaluating Memory of LLM-based Personal Assistants

  • Zeyu Zhang
  • Quanyu Dai
  • Luyu Chen
  • Zeren Jiang
  • Rui Li
  • Jieming Zhu
  • Xu Chen
  • Yi Xie

LLM-based agents have been widely applied as personal assistants, capable of memorizing information from user messages and responding to personal queries. However, there still lacks an objective and automatic evaluation on their memory capability, largely due to the challenges in constructing reliable questions and answers (QAs) according to user messages. In this paper, we propose MemSim, a Bayesian simulator designed to automatically construct reliable QAs from generated user messages, simultaneously keeping their diversity and scalability. Specifically, we introduce the Bayesian Relation Network (BRNet) and a causal generation mechanism to mitigate the impact of LLM hallucinations on factual information, facilitating the automatic creation of an evaluation dataset. Based on MemSim, we generate a dataset in the daily-life scenario, named MemDaily, and conduct extensive experiments to assess the effectiveness of our approach. We also provide a benchmark for evaluating different memory mechanisms in LLM-based agents with the MemDaily dataset.

PDF Details

AAAI Conference 2025 Conference Paper

MSSDA: Multi-Sub-Source Domain Adaptation for Diabetic Foot Neuropathy Recognition

  • Yan Zhong
  • Zhixin Yan
  • Yi Xie
  • Shibin Wu
  • Huaidong Zhang
  • Lin Shu
  • Peiru Zhou

Diabetic foot neuropathy (DFN) is a critical factor leading to diabetic foot ulcers, which is one of the most common and severe complications of diabetes mellitus (DM) and is associated with high risks of amputation and mortality. Despite its significance, existing datasets do not directly derive from plantar data and lack continuous, long-term foot-specific information. To advance DFN research, we have collected a novel dataset comprising continuous plantar pressure data to recognize diabetic foot neuropathy. This dataset includes data from 94 DM patients with DFN and 41 DM patients without DFN. Moreover, traditional methods divide datasets by individuals, potentially leading to significant domain discrepancies in some feature spaces due to the absence of mid-domain data. In this paper, we propose an effective domain adaptation method to address this proplem. We split the dataset based on convolutional feature statistics and select appropriate sub-source domains to enhance efficiency and avoid negative transfer. We then align the distributions of each source and target domain pair in specific feature spaces to minimize the domain gap. Comprehensive results validate the effectiveness of our method on both the newly proposed dataset for DFN recognition and an existing dataset.

PDF Details DOI

AAAI Conference 2024 Conference Paper

SAME: Sample Reconstruction against Model Extraction Attacks

  • Yi Xie
  • Jie Zhang
  • Shiqian Zhao
  • Tianwei Zhang
  • Xiaofeng Chen

While deep learning models have shown significant performance across various domains, their deployment needs extensive resources and advanced computing infrastructure. As a solution, Machine Learning as a Service (MLaaS) has emerged, lowering the barriers for users to release or productize their deep learning models. However, previous studies have highlighted potential privacy and security concerns associated with MLaaS, and one primary threat is model extraction attacks. To address this, there are many defense solutions but they suffer from unrealistic assumptions and generalization issues, making them less practical for reliable protection. Driven by these limitations, we introduce a novel defense mechanism, SAME, based on the concept of sample reconstruction. This strategy imposes minimal prerequisites on the defender's capabilities, eliminating the need for auxiliary Out-of-Distribution (OOD) datasets, user query history, white-box model access, and additional intervention during model training. It is compatible with existing active defense methods. Our extensive experiments corroborate the superior efficacy of SAME over state-of-the-art solutions. Our code is available at https://github.com/xythink/SAME.

PDF Details DOI

AAAI Conference 2023 Conference Paper

Easy Begun Is Half Done: Spatial-Temporal Graph Modeling with ST-Curriculum Dropout

  • Hongjun Wang
  • Jiyuan Chen
  • Tong Pan
  • Zipei Fan
  • Xuan Song
  • Renhe Jiang
  • Lingyu Zhang
  • Yi Xie

Spatial-temporal (ST) graph modeling, such as traffic speed forecasting and taxi demand prediction, is an important task in deep learning area. However, for the nodes in the graph, their ST patterns can vary greatly in difficulties for modeling, owning to the heterogeneous nature of ST data. We argue that unveiling the nodes to the model in a meaningful order, from easy to complex, can provide performance improvements over traditional training procedure. The idea has its root in Curriculum Learning, which suggests in the early stage of training models can be sensitive to noise and difficult samples. In this paper, we propose ST-Curriculum Dropout, a novel and easy-to-implement strategy for spatial-temporal graph modeling. Specifically, we evaluate the learning difficulty of each node in high-level feature space and drop those difficult ones out to ensure the model only needs to handle fundamental ST relations at the beginning, before gradually moving to hard ones. Our strategy can be applied to any canonical deep learning architecture without extra trainable parameters, and extensive experiments on a wide range of datasets are conducted to illustrate that, by controlling the difficulty level of ST relations as the training progresses, the model is able to capture better representation of the data and thus yields better generalization.

PDF Details DOI

ECAI Conference 2023 Conference Paper

Identifying Helpful Learnwares Without Examining the Whole Market

  • Yi Xie
  • Zhi-Hao Tan
  • Yuan Jiang 0001
  • Zhi-Hua Zhou

The learnware paradigm aims to construct a market of numerous well-performing machine learning models, which enables users to leverage these models to accomplish specific tasks without having to build models from scratch. Each learnware in the market is a model associated with a specification, representing the model’s utility and enabling it to be identified according to future users’ requirements. In the learnware paradigm, due to the vast and ever-increasing number of models in the market, a significant challenge is to identify helpful learnwares efficiently for a specific user task without leaking data privacy. However, existing identification methods require examining the whole market, which is computationally unaffordable in a large market. In this paper, we propose a new framework for identifying helpful learnwares without examining the whole market. Specifically, using the Reduced Kernel Mean Embedding (RKME) specification, we derive a novel learnware scoring criterion for assessing the helpfulness of a learnware, based on which we design an anchor-based framework to identify helpful learnwares by examining only a small portion of learnwares in the market. Theoretical analyses are provided for both the criterion and the anchor-based method. Empirical studies on market containing thousands of learnwares from real-world datasets confirm the effectiveness of our proposed approach.

Details

ICML Conference 2023 Conference Paper

Lookahead When It Matters: Adaptive Non-causal Transformers for Streaming Neural Transducers

  • Grant P. Strimel
  • Yi Xie
  • Brian John King
  • Martin Radfar
  • Ariya Rastrow
  • Athanasios Mouchtaris

Streaming speech recognition architectures are employed for low-latency, real-time applications. Such architectures are often characterized by their causality. Causal architectures emit tokens at each frame, relying only on current and past signal, while non-causal models are exposed to a window of future frames at each step to increase predictive accuracy. This dichotomy amounts to a trade-off for real-time Automatic Speech Recognition (ASR) system design: profit from the low-latency benefit of strictly-causal architectures while accepting predictive performance limitations, or realize the modeling benefits of future-context models accompanied by their higher latency penalty. In this work, we relax the constraints of this choice and present the Adaptive Non-Causal Attention Transducer (ANCAT). Our architecture is non-causal in the traditional sense, but executes in a low-latency, streaming manner by dynamically choosing when to rely on future context and to what degree within the audio stream. The resulting mechanism, when coupled with our novel regularization algorithms, delivers comparable accuracy to non-causal configurations while improving significantly upon latency, closing the gap with their causal counterparts. We showcase our design experimentally by reporting comparative ASR task results with measures of accuracy and latency on both publicly accessible and production-scale, voice-assistant datasets.

Details

EAAI Journal 2022 Journal Article

An adaptive decoding biased random key genetic algorithm for cloud workflow scheduling

  • Yi Xie
  • Yuhan Sheng
  • Moqi Qiu
  • Fengxian Gui

With the ever-growing data and computing requirements, more and more scientific and business applications represented by workflows have been moved or are in active transition to cloud platforms. Therefore, the cloud workflow scheduling has become a hot topic. As a well-known NP-hard problem, many heuristic or metaheuristic algorithms/methods have been proposed. However, the heuristic method is problem-dependent which fits only a particular of problems, while the metaheuristic method has the problems of incomplete search space or low search efficiency in the complete space. To fill these gaps, a novel adaptive decoding biased random key genetic algorithm for cloud workflow scheduling is proposed. In this algorithm, the improved real number coding based on random key with limited value range is employed, and some novel schemes such as the population initialization based on level and heuristics including dynamic heterogeneous earliest finish time, the dynamic adaptive decoding, the load balance with communication avoidance and iterative forward–backward scheduling are designed for population initialization, chromosome decoding and improvement. To evaluate the performance, extensive experiments have been conducted on various real and random workflow applications, which demonstrates that the proposed algorithm outperforms the conventional approaches.

Details DOI

NeurIPS Conference 2022 Conference Paper

Real-Valued Backpropagation is Unsuitable for Complex-Valued Neural Networks

  • Zhi-Hao Tan
  • Yi Xie
  • Yuan Jiang
  • Zhi-Hua Zhou

Recently complex-valued neural networks have received increasing attention due to successful applications in various tasks and the potential advantages of better theoretical properties and richer representational capacity. However, the training dynamics of complex networks compared to real networks remains an open problem. In this paper, we investigate the dynamics of deep complex networks during real-valued backpropagation in the infinite-width limit via neural tangent kernel (NTK). We first extend the Tensor Program to the complex domain, to show that the dynamics of any basic complex network architecture is governed by its NTK under real-valued backpropagation. Then we propose a way to investigate the comparison of training dynamics between complex and real networks by studying their NTKs. As a result, we surprisingly prove that for most complex activation functions, the commonly used real-valued backpropagation reduces the training dynamics of complex networks to that of ordinary real networks as the widths tend to infinity, thus eliminating the characteristics of complex-valued neural networks. Finally, the experiments validate our theoretical findings numerically.

PDF Details

NeurIPS Conference 2021 Conference Paper

CHIP: CHannel Independence-based Pruning for Compact Neural Networks

  • Yang Sui
  • Miao Yin
  • Yi Xie
  • Huy Phan
  • Saman Aliari Zonouz
  • Bo Yuan

Filter pruning has been widely used for neural network compression because of its enabled practical acceleration. To date, most of the existing filter pruning works explore the importance of filters via using intra-channel information. In this paper, starting from an inter-channel perspective, we propose to perform efficient filter pruning using Channel Independence, a metric that measures the correlations among different feature maps. The less independent feature map is interpreted as containing less useful information$/$knowledge, and hence its corresponding filter can be pruned without affecting model capacity. We systematically investigate the quantification metric, measuring scheme and sensitiveness$/$reliability of channel independence in the context of filter pruning. Our evaluation results for different models on various datasets show the superior performance of our approach. Notably, on CIFAR-10 dataset our solution can bring $0. 75\%$ and $0. 94\%$ accuracy increase over baseline ResNet-56 and ResNet-110 models, respectively, and meanwhile the model size and FLOPs are reduced by $42. 8\%$ and $47. 4\%$ (for ResNet-56) and $48. 3\%$ and $52. 1\%$ (for ResNet-110), respectively. On ImageNet dataset, our approach can achieve $40. 8\%$ and $44. 8\%$ storage and computation reductions, respectively, with $0. 15\%$ accuracy increase over the baseline ResNet-50 model. The code is available at https: //github. com/Eclipsess/CHIP_NeurIPS2021.

PDF Details

AAAI Conference 2021 Conference Paper

Enabling Fast and Universal Audio Adversarial Attack Using Generative Model

  • Yi Xie
  • Zhuohang Li
  • Cong Shi
  • Jian Liu
  • Yingying Chen
  • Bo Yuan

Recently, the vulnerability of deep neural network (DNN)based audio systems to adversarial attacks has obtained increasing attention. However, the existing audio adversarial attacks allow the adversary to possess the entire user’s audio input as well as granting sufficient time budget to generate the adversarial perturbations. These idealized assumptions, however, make the existing audio adversarial attacks mostly impossible to be launched in a timely fashion in practice (e. g. , playing unnoticeable adversarial perturbations along with user’s streaming input). To overcome these limitations, in this paper we propose fast audio adversarial perturbation generator (FAPG), which uses generative model to generate adversarial perturbations for the audio input in a single forward pass, thereby drastically improving the perturbation generation speed. Built on the top of FAPG, we further propose universal audio adversarial perturbation generator (UAPG), a scheme to craft universal adversarial perturbation that can be imposed on arbitrary benign audio input to cause misclassification. Extensive experiments on DNN-based audio systems show that our proposed FAPG can achieve high success rate with up to 214× speedup over the existing audio adversarial attack methods. Also our proposed UAPG generates universal adversarial perturbations that can achieve much better attack performance than the state-of-the-art solutions.

PDF Details

IJCAI Conference 2021 Conference Paper

Graph Universal Adversarial Attacks: A Few Bad Actors Ruin Graph Learning Models

  • Xiao Zang
  • Yi Xie
  • Jie Chen
  • Bo Yuan

Deep neural networks, while generalize well, are known to be sensitive to small adversarial perturbations. This phenomenon poses severe security threat and calls for in-depth investigation of the robustness of deep learning models. With the emergence of neural networks for graph structured data, similar investigations are urged to understand their robustness. It has been found that adversarially perturbing the graph structure and/or node features may result in a significant degradation of the model performance. In this work, we show from a different angle that such fragility similarly occurs if the graph contains a few bad-actor nodes, which compromise a trained graph neural network through flipping the connections to any targeted victim. Worse, the bad actors found for one graph model severely compromise other models as well. We call the bad actors ``anchor nodes'' and propose an algorithm, named GUA, to identify them. Thorough empirical investigations suggest an interesting finding that the anchor nodes often belong to the same class; and they also corroborate the intuitive trade-off between the number of anchor nodes and the attack success rate. For the dataset Cora which contains 2708 nodes, as few as six anchor nodes will result in an attack success rate higher than 80% for GCN and other three models.

PDF Details DOI

AAAI Conference 2020 Conference Paper

CAG: A Real-Time Low-Cost Enhanced-Robustness High-Transferability Content-Aware Adversarial Attack Generator

  • Huy Phan
  • Yi Xie
  • Siyu Liao
  • Jie Chen
  • Bo Yuan

Deep neural networks (DNNs) are vulnerable to adversarial attack despite their tremendous success in many artificial intelligence fields. Adversarial attack is a method that causes the intended misclassfication by adding imperceptible perturbations to legitimate inputs. To date, researchers have developed numerous types of adversarial attack methods. However, from the perspective of practical deployment, these methods suffer from several drawbacks such as long attack generating time, high memory cost, insufficient robustness and low transferability. To address the drawbacks, we propose a Content-aware Adversarial Attack Generator (CAG) to achieve real-time, low-cost, enhanced-robustness and hightransferability adversarial attack. First, as a type of generative model-based attack, CAG shows significant speedup (at least 500 times) in generating adversarial examples compared to the state-of-the-art attacks such as PGD and C&W. Furthermore, CAG only needs a single generative model to perform targeted attack to any targeted class. Because CAG encodes the label information into a trainable embedding layer, it differs from prior generative model-based adversarial attacks that use n different copies of generative models for n different targeted classes. As a result, CAG significantly reduces the required memory cost for generating adversarial examples. Moreover, CAG can generate adversarial perturbations that focus on the critical areas of input by integrating the class activation maps information in the training process, and hence improve the robustness of CAG attack against the state-of-art adversarial defenses. In addition, CAG exhibits high transferability across different DNN classifier models in black-box attack scenario by introducing random dropout in the process of generating perturbations. Extensive experiments on different datasets and DNN models have verified the realtime, low-cost, enhanced-robustness, and high-transferability benefits of CAG.

PDF Details
v2026.09.13