人工智能韧性研究现状及展望
Artificial Intelligence Resilience: Current State and Future Perspectives
人工智能(AI)技术正深度融入关键基础设施,其韧性对保障系统安全稳定运行至关重要。本文将AI韧性定义为稳健性、防御力、复原力和进化力4个核心维度,系统梳理AI韧性研究的现状,围绕上述4个核心维度综述国内外关键技术进展,并特别关注大语言模型等新技术带来的新挑战与新方案。在此基础上,研究提出了当前AI韧性发展面临的能力建设缺乏顶层规划、评测体系缺少真实场景、大模型韧性重视不足等突出问题。研究建议:加强战略引领,构建系统化韧性框架;建设高保真、多维度、可复现的韧性评测体系;重点挖掘大模型潜力,推动其在“训练 ‒ 部署 ‒ 运行 ‒ 更新”全生命周期的多层级韧性能力提升,以构建更可靠、可信且持续的智能系统。
Artificial intelligence (AI) technologies are being deeply integrated into critical infrastructures, making AI resilience essential to ensuring the secure and stable operation of such systems. This study defines AI resilience in terms of four core dimensions—robustness, defensibility, recoverability, and evolvability—and reviews the current state of research in this area. Focusing on these four dimensions, we survey key technical advances both in China and abroad, with particular attention to new challenges and emerging solutions brought about by technologies such as large language models (LLMs). On this basis, we identify several prominent issues hindering the development of AI resilience, including the lack of top-level planning for capability building, absence of evaluation frameworks grounded in realistic application scenarios, and insufficient emphasis on the resilience of LLMs. To address these challenges, we recommend strengthening strategic guidance to establish a systematic resilience framework; developing high-fidelity, multi-dimensional, and reproducible evaluation systems; and exploring the potentials of LLMs to enhance multi-level resilience across the entire lifecycle of training, deployment, operation, and update, thereby enabling the construction of more reliable, trustworthy, and sustainable intelligent systems.
| [1] |
Tran T, Pham T, Carneiro G, et al. A bayesian data augmentation approach for learning deep models [R]. Long Beach: The 2017 Conference on Neural Information Processing Systems, 2017. |
| [2] |
Dong J H, Cong Y, Sun G, et al. Where and how to transfer: Knowledge aggregation-induced transferability perception for unsupervised domain adaptation [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024, 46(3): 1664‒1681. |
| [3] |
Gao Y S, Xu C, Wang D R, et al. STRIP: A defence against trojan attacks on deep neural networks [R]. San Juan: The 35th Annual Computer Security Applications Conference, 2019. |
| [4] |
Han D Q, Wang Z L, Zhong Y, et al. Evaluating and improving adversarial robustness of machine learning-based network intrusion detectors [J]. IEEE Journal on Selected Areas in Communications, 2021, 39(8): 2632‒2647. |
| [5] |
Zhang J, Chen X, Xiang Y, et al. Robust network traffic classification [J]. IEEE/ACM Transactions on Networking, 2015, 23(4): 1257‒1270. |
| [6] |
Liu C, He L T, Xiong G, et al. FS-net: A flow sequence network for encrypted traffic classification [R]. Paris: IEEE INFOCOM 2019—IEEE Conference on Computer Communications, 2019. |
| [7] |
Goodfellow I, Pouget-Abadie J, Mirza M, et al. Generative adversarial networks [J]. Conmmunications of the ACM, 2020, 63(11): 139‒144. |
| [8] |
Shmakov A, Greif K, Fenton M, et al. End-to-end latent variational diffusion models for inverse problems in high energy physics [R]. New Orleans: The 37th International Conference on Neural Information Processing Systems, 2023. |
| [9] |
Wang Z K, Pang T Y, Du C, et al. Better diffusion models further improve adversarial training [R]. Honolulu: The 40th International Conference on Machine Learning, 2023. |
| [10] |
Chawla N V, Bowyer K W, Hall L O, et al. SMOTE: Synthetic minority over-sampling technique [J]. Journal of Artificial Intelligence Research, 2002, 16: 321‒357. |
| [11] |
Tobin J, Fong R, Ray A, et al. Domain randomization for transferring deep neural networks from simulation to the real world [R]. Vancouver: 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2017. |
| [12] |
Peng X B, Andrychowicz M, Zaremba W, et al. Sim-to-real transfer of robotic control with dynamics randomization [R]. Brisbane: 2018 IEEE International Conference on Robotics and Automation (ICRA), 2018. |
| [13] |
Lundberg S M, Lee S I. A unified approach to interpreting model predictions [R]. Long Beach: The 31st International Conference on Neural Information Processing Systems, 2017. |
| [14] |
Badreddine S, d'Avila Garcez A, Serafini L, et al. Logic tensor networks [J]. Artificial Intelligence, 2022, 303: 103649. |
| [15] |
Selvaraju R R, Cogswell M, Das A, et al. Grad-CAM: Visual explanations from deep networks via gradient-based localization [J]. International Journal of Computer Vision, 2020, 128(2): 336‒359. |
| [16] |
Liu Y A, Zhang R Q, Zhang M K, et al. Perturbation-invariant adversarial training for neural ranking models: Improving the effectiveness‒robustness trade-off [R]. Vancouver: The 38th Annual the AAAI Conference on Artificial Intelligence, 2024, 38(8): 8832‒8840. |
| [17] |
Yao L F, Niu W N, Yuan Q J, et al. A robust malicious traffic detection framework with low-quality labeled data [R]. Denver: ICC 2024‒IEEE International Conference on Communications, 2024. |
| [18] |
Han B, Yao Q, Yu X, et al. Co-teaching: Robust training of deep neural networks with extremely noisy labels [R]. Montréal: The 2018 Conference on Neural Information Processing Systems, 2018. |
| [19] |
Yuan Q J, Gou G P, Zhu Y B, et al. MCRe: A unified framework for handling malicious traffic with noise labels based on multidimensional constraint representation [J]. IEEE Transactions on Information Forensics and Security, 2024, 19: 133‒147. |
| [20] |
Li Y, Lyu X, Koren N, et al. Anti-backdoor learning: Training clean models on poisoned data [R]. Virtual: The 2021 Conference on Neural Information Processing Systems, 2021. |
| [21] |
Li T, Li H E, Pan Y N, et al. Meta stackelberg game: Robust federated learning against adaptive and mixed poisoning attacks [EB/OL]. (2024-10-22)[2025-03-10]. https://arxiv.org/abs/2410.17431. |
| [22] |
Madry A, Makelov A, Schmidt L, et al. Towards deep learning models resistant to adversarial attacks [R]. Vancouver: International Conference on Learning Representations, 2018. |
| [23] |
Pinto L, Davidson J, Sukthankar R, et al. Robust adversarial reinforcement learning [R]. Sydney: The 34th International Conference on Machine Learning (ICML 2017), 2017. |
| [24] |
Kalashnikov D, Irpan A, Pastor P, et al. Scalable deep reinforcement learning for vision-based robotic manipulation [R]. Zürich: The 2nd Annual Conference on Robot Learning (CoRL 2018), 2018. |
| [25] |
Ye J J, Wu Y L, Gao S Y, et al. RoTBench: A multi-level benchmark for evaluating the robustness of large language models in tool learning [R]. Miami: The 2024 Conference on Empirical Methods in Natural Language Processing, 2024. |
| [26] |
Huang Y, Sun L, Wang H, et al. Position: TrustLLM: Trustworthiness in large language models [R]. Baltimore: The 41st International Conference on Machine Learning (ICML 2024), 2024. |
| [27] |
Chang Y P, Wang X, Wang J D, et al. A survey on evaluation of large language models [J]. ACM Transactions on Intelligent Systems and Technology, 2024, 15(3): 1‒45. |
| [28] |
Liu Y, Yao Y S, Ton J F, et al. Trustworthy LLMs: A survey and guideline for evaluating large language models' alignment [EB/OL]. (2023-08-10)[2025-03-02]. https://arxiv.org/abs/2308.05374. |
| [29] |
Qiu H C, Zhang S, Li A Q, et al. Latent jailbreak: A benchmark for evaluating text safety and output robustness of large language models [EB/OL]. (2023-07-17)[2025-03-02]. https://arxiv.org/abs/2307.08487. |
| [30] |
Dobre D, Gidel G, Günnemann S, et al. Soft prompt threats: Attacking safety alignment and unlearning in open-source LLMs through the embedding space [R]. Vancouver: Advances in Neural Information Processing Systems 37 (NeuralIPS 2024), 2024. |
| [31] |
Luo W D, Ma S Y, Liu X G, et al. JailBreakV: A benchmark for assessing the robustness of MultiModal large language models against jailbreak attacks [EB/OL]. (2024-04-03)[2025-03-06]. https://arxiv.org/abs/2404.03027. |
| [32] |
Zhao X D, Yang X J, Pang T Y, et al. Weak-to-strong jailbreaking on large language models [EB/OL]. (2024-01-30)[2025-03-12]. https://arxiv.org/abs/2401.17256. |
| [33] |
Ouyang L, Wu J, Jiang X, et al. Training language models to follow instructions with human feedback [R]. New Orleans: The 2022 Conference on Neural Information Processing Systems, 2022. |
| [34] |
Dettmers T, Pagnoni A, Holtzman A, et al. QLoRA: Efficient finetuning of quantized LLMs [R]. New Orleans: the 2023 Conference on Neural Information Processing Systems, 2023. |
| [35] |
Rombach R, Blattmann A, Lorenz D, et al. High-resolution image synthesis with latent diffusion models [R]. New Orleans: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022. |
| [36] |
Wang S B, Zhang J, Yuan Z, et al. Pre-trained model guided fine-tuning for zero-shot adversarial robustness [R]. Seattle: 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024. |
| [37] |
Zhang H H, Kung P N, Yoshida M, et al. Adaptable logical control for large language models [R]. Vancouver: The 38th International Conference on Neural Information Processing Systems (NeurIPS 2024), 2024. |
| [38] |
Abadi M, Chu A, Goodfellow I, et al. Deep learning with differential privacy [R]. Vienna: The 2016 ACM SIGSAC Conference on Computer and Communications Security, 2016. |
| [39] |
Wang X L, Jin Y C. Distilling ensemble surrogates for federated data-driven many-task optimization [J]. IEEE Transactions on Evolutionary Computation, 2025, 29(6): 2401‒2415. |
| [40] |
Tan Q, Li Q, Zhao Y, et al. Defending against data reconstruction attacks in federated learning: An information theory approach [R]. Philadelphia: The 33rd USENIX Conference on Security Symposium, 2024. |
| [41] |
Xu Y X, Fang B X, Li M H, et al. Query-efficient model inversion attacks: An information flow view [J]. IEEE Transactions on Information Forensics and Security, 2025, 20: 1023‒1036. |
| [42] |
Tan T J L, Shokri R. Bypassing backdoor detection algorithms in deep learning [R]. Genoa: 2020 IEEE European Symposium on Security and Privacy (EuroS&P), 2020. |
| [43] |
Steinhardt J, Koh P W, Liang P. Certified defenses for data poisoning attacks [R]. Long Beach: The 30th International Conference on Neural Information Processing Systems (NeurIPS 2017), 2017. |
| [44] |
Huang H, Erfani S M, Li Y, et al. Detecting backdoor samples in contrastive language image pretraining [R]. Singapore: The 13th International Conference on Learning Representations (ICLR 2025), 2025. |
| [45] |
Wang N, Chen Y M, Hu Y, et al. MANDA: On adversarial example detection for network intrusion detection system [R]. Vancouver: IEEE INFOCOM 2021—IEEE Conference on Computer Communications, 2021. |
| [46] |
Jiang W B, Li H W, Xu G W, et al. A comprehensive defense framework against model extraction attacks [J]. IEEE Transactions on Dependable and Secure Computing, 2024, 21(2): 685‒700. |
| [47] |
Wang B L, Yao Y S, Shan S, et al. Neural cleanse: Identifying and mitigating backdoor attacks in neural networks [R]. San Francisco: 2019 IEEE Symposium on Security and Privacy (SP), 2019. |
| [48] |
Liu X G, Li M H, Wang H Y, et al. Detecting backdoors during the inference stage based on corruption robustness consistency [R]. Vancouver: 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023. |
| [49] |
Fang B X, Li M H, Tang K K, et al. LT-defense: Searching-free backdoor defense via exploiting the long-tailed effect [R]. Vancouver: Advances in Neural Information Processing Systems 37, 2024. |
| [50] |
Tramèr F, Boneh D. Adversarial training and robustness for multiple perturbations [R]. Red Hook: The 32th International Conference on Neural Information Processing Systems (NeurIPS 2024), 2019. |
| [51] |
Qi F C, Chen Y Y, Li M K, et al. ONION: A simple and effective defense against textual backdoor attacks [R]. Punta Cana: The 2021 Conference on Empirical Methods in Natural Language Processing, 2021. |
| [52] |
Shan S, Wenger E, Wang B L, et al. Gotta Catch'Em all: Using honeypots to catch adversarial attacks on neural networks [R]. Online: The 2020 ACM SIGSAC Conference on Computer and Communications Security, 2020. |
| [53] |
Gong X L, Wang Z Y, Li S K, et al. A GAN-based defense framework against model inversion attacks [J]. IEEE Transactions on Information Forensics and Security, 2023, 18: 4475‒4487. |
| [54] |
Xu Y X, Fang B X, Wang R, et al. Neural honeytrace: A robust plug-and-play watermarking framework against model extraction attacks [EB/OL]. (2025-01-16)[2025-03-012]. https://arxiv.org/abs/2501.09328. |
| [55] |
Gao L, Fu C P, Deng X H, et al. Wedjat: Detecting sophisticated evasion attacks via real-time causal analysis [R]. Toronto: The 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.1, 2025. |
| [56] |
Zhu H, Zhang S Z, Chen K. AI-guardian: Defeating adversarial attacks using backdoors [R]. San Francisco: 2023 IEEE Symposium on Security and Privacy (SP), 2023. |
| [57] |
Feng R, Hooda A, Mangaokar N, et al. Stateful defenses for machine learning models are not yet secure against black-box attacks [R]. Copenhagen: The 2023 ACM SIGSAC Conference on Computer and Communications Security, 2023. |
| [58] |
Zhang X Y, Zhang C, Li T L, et al. JailGuard: A universal detection framework for prompt-based attacks on LLM systems [J]. ACM Transactions on Software Engineering and Methodology, 2025, 35(1): 1‒40. |
| [59] |
Schramowski P, Brack M, Deiseroth B, et al. Safe latent diffusion: Mitigating inappropriate degeneration in diffusion models [R]. Vancouver: 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023. |
| [60] |
台建玮, 杨双宁, 王佳佳, 大语言模型对抗性攻击与防御综述 [J]. 计算机研究与发展, 2025, 62(3): 563‒588. |
| [61] |
Tai J W, Yang S N, Wang J J, et al. Survey of adversarial attacks and defenses for large language models [J]. Journal of Computer Research and Development, 2025, 62(3): 563‒588. |
| [62] |
Li N, Han Z W, Steneker I, et al. LLM defenses are not robust to multi-turn human jailbreaks yet [R]. Vancouver: Red Teaming GenAI Workshop, The 38th International Conference on Neural Information Processing Systems (NeurIPS 2024), 2024. |
| [63] |
Jain N, Schwarzschild A, Wen Y X, et al. Baseline defenses for adversarial attacks against aligned language models [EB/OL]. (2023-09-01)[2025-03-25]. https://arxiv.org/abs/2309.00614. |
| [64] |
Kawasaki A, Davis A, Abbas H. Defending large language models against attacks with residual stream activation analysis [R]. Arlington: The Conference on Applied Machine Learning in Information Security (CAMLIS 2024), 2024. |
| [65] |
Xie Y Q, Fang M H, Pi R J, et al. GradSafe: Detecting jailbreak prompts for LLMs via safety-critical gradient analysis [R]. Bangkok: The 62nd Annual Meeting of the Association for Computational Linguistics, 2024. |
| [66] |
Zhang H F, Guo Z M, Zhu H S, et al. Jailbreak open-sourced large language models via enforced decoding [R]. Bangkok: The 62nd Annual Meeting of the Association for Computational Linguistics, 2024. |
| [67] |
Ji J B, Hou B R, Robey A, et al. Defending large language models against jailbreak attacks via semantic smoothing [EB/OL]. (2024-02-25)[2025-03-25]. https://arxiv.org/abs/2402.16192. |
| [68] |
Robey A, Wong E, Hassani H, et al. SmoothLLM: Defending large language models against jailbreaking attacks [EB/OL]. (2023-10-05)[2025-07-25]. https://arxiv.org/abs/2310.03684. |
| [69] |
Zhao W, Li Z, Li Y G, et al. Defending large language models against jailbreak attacks via layer-specific editing [R]. Miami: The 2024 Conference on Empirical Methods in Natural Language Processing, 2024. |
| [70] |
Tang R, Yuan J, Li Y, et al. Setting the Trap: Capturing and defeating backdoors in pretrained language models through honeypots [R]. Orleans: The 37th International Conference on Neural Information Processing Systems (NeurIPS 2023), 2023. |
| [71] |
Liu Q, Yin J, Wen W, et al. NeuroPots: Realtime proactive defense against bit-flip attacks in neural networks [R]. Anaheim: The 32nd USENIX Security Symposium (USENIX Security 2023), 2023. |
| [72] |
Andreina S, Marson G A, Möllering H, et al. BaFFLe: Backdoor detection via feedback-based federated learning [R]. Washington DC: 2021 IEEE 41st International Conference on Distributed Computing Systems (ICDCS), 2021. |
| [73] |
Bongard J, Zykov V, Lipson H. Resilient machines through continuous self-modeling [J]. Science, 2006, 314(5802): 1118‒1121. |
| [74] |
Cully A, Clune J, Tarapore D, et al. Robots that can adapt like animals [J]. Nature, 2015, 521(7553): 503‒507. |
| [75] |
Zhao S, Tuan L A, Fu J, et al. Exploring clean label backdoor attacks and defense in language models [J]. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2024, 32: 3014‒3024. |
| [76] |
Huang H Y, Wang Q, Gong X L, et al. Orion: Online backdoor sample detection via evolution deviance [R]. Macao: The Thirty-Second International Joint Conference on Artificial Intelligence, 2023. |
| [77] |
Liu K, Dolan-Gavitt B, Garg S. Fine-pruning: Defending against backdooring attacks on deep neural networks [R]. Cham: Research in Attacks, Intrusions, and Defenses, 2018. |
| [78] |
Guo W B, Wang L, Xu Y, et al. Towards inspecting and eliminating Trojan backdoors in deep neural networks [R]. Sorrento: 2020 IEEE International Conference on Data Mining (ICDM), 2020. |
| [79] |
Zhu L W, Ning R, Wang C, et al. GangSweep: Sweep out neural backdoors by GAN [R]. Seattle: The 28th ACM International Conference on Multimedia, 2020. |
| [80] |
Wu D, Wang Y. Adversarial neuron pruning purifies backdoored deep models [R]. Online: The 34th International Conference on Neural Information Processing Systems (NeurIPS 2021), 2021. |
| [81] |
Li Y G, Lyu X X, Koren N, et al. Neural attention distillation: Erasing backdoor triggers from deep neural networks [R]. Online: The 9th International Conference on Learning Representations (ICLR 2021), 2021. |
| [82] |
Xu Q L, Tao G H, Honorio J, et al. MEDIC: Remove model backdoors via importance driven cloning [R]. Vancouver: 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023. |
| [83] |
Liu Y, Fan M Y, Chen C, et al. Backdoor defense with machine unlearning [R]. London: IEEE INFOCOM 2022—IEEE Conference on Computer Communications, 2022. |
| [84] |
Chen X, Ding K, Fan C Y, et al. Defensive unlearning with adversarial training for robust concept erasure in diffusion models [R]. Vancouver: The 38th International Conference on Neural Information Processing Systems (NeurIPS 2024), 2024. |
| [85] |
Sun B, Sun J, Pham L H, et al. Causality-based neural network repair [R]. Pittsburgh: 2022 IEEE/ACM 44th International Conference on Software Engineering (ICSE), 2022. |
| [86] |
Zhang T, Zhang W J, Gupta M M, et al. Resilient robots: Concept, review, and future directions [J]. Robotics, 2017, 6(4): 22‒36. |
| [87] |
Chen S, Li B. Toward adaptive reasoning in large language models with thought rollback [R]. Vienna: The 41st International Conference on Machine Learning (ICML 2024), 2024. |
| [88] |
Lo M, Barez F, Cohen S. Large language models relearn removed concepts [R]. Bangkok: The 62nd Annual Meeting of the Association for Computational Linguistics ACL 2024, 2024. |
| [89] |
Zhang L, Jin Q, Huang H Y, et al. Respond in my language: Mitigating language inconsistency in response generation based on large language models [R]. Bangkok: The 62nd Annual Meeting of the Association for Computational Linguistics, 2024. |
| [90] |
Zhang X Y, Peng B L, Tian Y, et al. Self-alignment for factuality: Mitigating hallucinations in LLMs via self-evaluation [R]. Bangkok: The 62nd Annual Meeting of the Association for Computational Linguistics, 2024. |
| [91] |
Zintgraf L M, Shiarlis K, Kurin V, et al. Fast context adaptation via meta-learning [R]. Long Beach: The 36th International Conference on Machine Learning (ICML 2019), 2019. |
| [92] |
Snell J, Swersky K, Zemel R S. Prototypical networks for few-shot learning [R]. Long Beach: The 30th International Conference on Neural Information Processing Systems (NeurIPS 2017), 2017. |
| [93] |
Sung F, Yang Y X, Zhang L, et al. Learning to compare: Relation network for few-shot learning [R]. Salt Lake City: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018. |
| [94] |
Jiang M H, Cui M X, Liu C, et al. Zero-relabelling mobile-app identification over drifted encrypted network traffic [J]. Computer Networks, 2023, 228: 109728. |
| [95] |
Pathak D, Agrawal P, Efros A A, et al. Curiosity-driven exploration by self-supervised prediction [R]. Sydney: The 34th International Conference on Machine Learning 2017, 2017. |
| [96] |
Pinto L, Gupta A. Supersizing self-supervision: Learning to grasp from 50K tries and 700 robot hours [R]. Stockholm: 2016 IEEE International Conference on Robotics and Automation (ICRA), 2016. |
| [97] |
Mallya A, Lazebnik S. PackNet: Adding multiple tasks to a single network by iterative pruning [R]. Salt Lake City: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018. |
| [98] |
Rusu A A, Rabinowitz N C, Desjardins G, et al. Progressive neural networks [EB/OL]. (2016-06-15)[2025-03-05]. https://arxiv.org/abs/1606.04671. |
| [99] |
Finn C, Abbeel P, Levine S. Model-agnostic meta-learning for fast adaptation of deep networks [R]. Sydney: The 34th International Conference on Machine Learning, 2017. |
| [100] |
Parisi G I, Kemker R, Part J L, et al. Continual lifelong learning with neural networks: A review [J]. Neural Networks, 2019, 113: 54‒71. |
| [101] |
Nagabandi A, Kahn G, Fearing R S, et al. Neural network dynamics for model-based deep reinforcement learning with model-free fine-tuning [R]. Brisbane: 2018 IEEE International Conference on Robotics and Automation (ICRA), 2018. |
| [102] |
Levine S, Finn C, Darrell T, et al. End-to-end training of deep visuomotor policies [J]. Journal of Machine Learning Research, 2016, 17(1): 1334‒1373. |
| [103] |
Zenke F, Poole B, Ganguli S. Continual learning through synaptic intelligence [R]. Sydney: The 34th International Conference on Machine Learning, 2017. |
| [104] |
Chen T, Kornblith S, Norouzi M, et al. A simple framework for contrastive learning of visual representations [R]. Vienna: The 37th International Conference on Machine Learning, 2020. |
| [105] |
He K M, Fan H Q, Wu Y X, et al. Momentum contrast for unsupervised visual representation learning [R]. Seattle: 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020. |
| [106] |
He K M, Chen X L, Xie S N, et al. Masked autoencoders are scalable vision learners [R]. New Orleans: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022. |
| [107] |
Xie R J, Wang Y X, Cao J H, et al. Rosetta: Enabling robust TLS encrypted traffic classification in diverse network environments with TCP-aware traffic augmentation [R]. Wuhan: The ACM Turing Award Celebration Conference—China 2023, 2023. |
| [108] |
Cai S H, Tang H, Chen J F, et al. CDDA-MD: An efficient malicious traffic detection method based on concept drift detection and adaptation technique [J]. Computers & Security, 2025, 148: 104121. |
| [109] |
Cheng S Y, Tao G H, Liu Y Q, et al. BEAGLE: Forensics of deep learning backdoor attack for better defense [R]. San Diego: The Network and Distributed System Security Symposium 2023, 2023. |
| [110] |
Ha D, Schmidhuber J. Recurrent world models facilitate policy evolution [R]. Montréal: The 32nd International Conference on Neural Information Processing Systems (NeurIPS 2018), 2018. |
| [111] |
Zhang Y Z, Abbeel P, Pinto L. Automatic curriculum learning through value disagreement [R]. Vancouver: The 34th International Conference on Neural Information Processing Systems (NeurIPS 2020), 2020. |
| [112] |
Zhao R, Song J M, Yuan Y F, et al. Maximum entropy population-based training for zero-shot human-AI coordination [J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2023, 37(5): 6145‒6153. |
| [113] |
Jang J, Ye S, Yang S, et al. Towards continual knowledge learning of language models [R]. Online: The International Conference on Learning Representations ICLR 2022, 2022. |
| [114] |
Hao S, Liu T, Wang Z, et al. Toolkengpt: Augmenting frozen language models with massive tools via tool embeddings [R]. New Orleans: The 37th International Conference on Neural Information Processing Systems (NeurIPS 2023), 2023. |
| [115] |
Lange R, Tian Y T, Tang Y J. Large language models as evolution strategies [R]. Melbourne: The Genetic and Evolutionary Computation Conference Companion, 2024. |
| [116] |
Du G D, Li J, Liu H T, et al. Knowledge fusion by evolving weights of language models [R]. Bangkok: The 2024 Conference on Empirical Methods in Natural Language Processing, 2024. |
| [117] |
Luo J Y, Zhang W Z, Yuan Y, et al. Large language model agent: A survey on methodology, applications and challenges [EB/OL]. (2025-03-27)[2025-04-28]. https://arxiv.org/abs/2503.21460. |
| [118] |
Dou Z Y, Yang C F, Wu X Q, et al. Re-ReST: Reflection-reinforced self-training for language agents [R]. Miami: The 2024 Conference on Empirical Methods in Natural Language Processing, 2024. |
| [119] |
Shinn N, Cassano F, Berman E, et al. Reflexion: Language agents with verbal reinforcement learning [R]. Orleans: The 37th International Conference on Neural Information Processing Systems (NeurIPS 2023), 2023. |
中国工程院咨询项目“关键信息基础设施网络韧性发展战略研究”(2023-JB-13)
国家自然科学基金项目(62372126)
国家自然科学基金项目(U2436208)
国家自然科学基金项目(62372129)
国家自然科学基金项目(U2468204)
广东省重点研发计划项目(2024B0101010002)
广东省工业控制系统攻防对抗重点实验室项目(2024B1212020010)
/
| 〈 |
|
〉 |