A journal of IEEE and CAA , publishes high-quality papers in English on original theoretical/experimental research and development in all areas of automation
Advanced Search

IEEE/CAA Journal of Automatica Sinica

  • JCR Impact Factor: 18.3, Top 1 (SCI Q1)
    CiteScore: 28.2, Top 1% (Q1)
    Google Scholar h5-index: 95， TOP 5
Turn off MathJax
Article Contents
Article Navigation >  IEEE/CAA Journal of Automatica Sinica  > 2026  > In Press, Accepted Manuscript
Z. Chen, Z. Ma, C. Yang, F. Li, Y. Liu, and W. Gui, IEEE/CAA J. Autom. Sinica, early access, 2026. doi: 10.1109/JAS.2026.126173
Citation: Z. Chen, Z. Ma, C. Yang, F. Li, Y. Liu, and W. Gui, IEEE/CAA J. Autom. Sinica, early access, 2026. doi: 10.1109/JAS.2026.126173
shu

MeMo: Embodied Intelligent Metallurgical Large Model for Non-ferrous Metals

doi: 10.1109/JAS.2026.126173
Funds:  This work was supported in part by the Key Program of the Joint Fund for Basic Research on Large Aircraft of the National Natural Science Foundation of China (U2570207), the Science and Technology Innovation Program of Hunan Province (2026RC1005), the Key Research and Development Program of Hunan Province (2025AQ2027), the Major Program of the National Natural Science Foundation of China (62394340), the National Natural Science Foundation of China (62273359, 62373377), the National Key Research and Development Program of China (2024YFC3908000, 2024YFC3908002), the Young Key Teachers Training Program of Hunan Province under Grant 206030802, and the Postgraduate Education Reform Program of Hunan Provincial Department of Education (2024JGYB021)
More Information
    Abstract
  • The development of large models is driving the exploration of AI in industrial applications. However, their black-box nature conflicts with the need for high reliability and interpretability in industrial production, limiting their practical use due to challenges in internalizing industrial knowledge and the absence of physical interaction. To address this challenge, we integrate the strengths of connectionism, symbolism, and behaviorism to develop an embodied intelligent control model for non-ferrous metallurgy. The model introduces a meta-model-driven fine-tuning strategy based on connectionism, converting noisy time-series data into meta-model code infused with expert knowledge. The question-answer-evaluation training framework guides the model from data imitation to a deeper understanding of industrial problems. Additionally, it incorporates a knowledge graph-embedded binarized reward model that converts process constraints into evaluation signals for the model, ensuring compliance with symbolic logic and industry laws while preventing reward hacking. Furthermore, a behaviorist virtual-real fusion feedback architecture is designed, enabling rapid task decomposition and verification on both virtual and physical platforms, while integrating feedback to optimize real-world performance. Experiments demonstrate that this approach enhances the generalization ability of large models on complex industrial data and enables optimization results that approach theoretical limits. This work establishes a new industrial intelligence paradigm that seamlessly integrates data, knowledge, and environmental feedback, offering a reliable path for deploying large models in the process industry.

     

    • FullText(HTML)
  • loading
    • References(36)
  • [1]
    A. Newell and H. Simon, “The logic theory machine–a complex information processing system,” IRE Trans. Inf. Theory, vol. 2, no. 3, pp. 61−79, Sep. 1956. doi: 10.7249/p868
    [2]
    J. A. Feldman and D. H. Ballard, “Connectionist models and their properties,” Cogn. Sci., vol. 6, no. 3, pp. 205−254, Jul. 1982. doi: 10.1207/s15516709cog0603_1
    [3]
    R. S. Sutton and A. G. Barto, Reinforcement Learning: An Introduction, vol. 1. Cambridge, MA, USA: MIT Press, 1998.
    [4]
    J. Tan, J. Shi, L. Wu, B. Chen, H. Tang, C. Zhang, W. Zhang, S. Wang, and J. Wan, “Embodied intelligence empowering customized manufacturing: Architecture, opportunities, and challenges,” IEEE Access, vol. 13, pp. 92740−92755, May 2025. doi: 10.1109/ACCESS.2025.3572778
    [5]
    H. Xue and F. D. Salim, “PromptCast: A new prompt-based learning paradigm for time series forecasting,” IEEE Trans. Knowl. Data Eng., vol. 36, no. 11, pp. 6851−6864, Nov. 2024. doi: 10.1109/TKDE.2023.3342137
    [6]
    D. Spathis and F. Kawsar, “The first step is the hardest: Pitfalls of representing and tokenizing temporal data for large language models,” J. Am. Med. Inform. Assoc., vol. 31, no. 9, pp. 2151−2158, Sep. 2024. doi: 10.1093/jamia/ocae090
    [7]
    Y. Duan, J. Zhou, Z. Wang, Y.-K. Wang, and C.-T. Lin, “DeWave: Discrete EEG waves encoding for brain dynamics to text translation,” in Proc. 37th Int. Conf. Neural Information Processing Systems, New Orleans, LA, USA, 2023, Art. no. 432.
    [8]
    D. Cao, F. Jia, S. Ö. Arik, T. Pfister, Y. Zheng, W. Ye, and Y. Liu, “TEMPO: Prompt-based generative pre-trained transformer for time series forecasting,” in Proc. 12th Int. Conf. Learning Representations, Vienna, Austria, 2024. (查阅网上资料, 未找到页码信息, 请确认补充)
    [9]
    M. Jin, S. Wang, L. Ma, Z. Chu, J. Y. Zhang, X. Shi, et al., “Time-LLM: Time series forecasting by reprogramming large language models,” in Proc. 12th Int. Conf. Learning Representations, Vienna, Austria, 2024. (查阅网上资料, 未找到页码信息, 请确认补充)
    [10]
    P. F. Christiano, J. Leike, T. B. Brown, M. Martic, S. Legg, and D. Amodei, “Deep reinforcement learning from human preferences,” in Proc. 31st Int. Conf. Neural Information Processing Systems, Long Beach, CA, USA, 2017, pp. 4302−4310.
    [11]
    M. Hwang, G. Lee, H. Kee, C. Kim, K. Lee, and S. Oh, “Sequential preference ranking for efficient reinforcement learning from human feedback,” in Proc. 37th Int. Conf. Neural Information Processing Systems, New Orleans, LA, USA, 2023, Art. no. 2133.
    [12]
    K. Lee, L. M. Smith, and P. Abbeel, “PEBBLE: Feedback-efficient interactive reinforcement learning via relabeling experience and unsupervised pre-training,” in Proc. 38th Int. Conf. Machine Learning, 2021, pp. 6152−6163. (查阅网上资料, 未找到对应的出版地信息, 请确认补充)
    [13]
    J. Park, Y. Seo, J. Shin, H. Lee, P. Abbeel, and K. Lee, “SURF: Semi-supervised reward learning with data augmentation for feedback-efficient preference-based reinforcement learning,” in Proc. 10th Int. Conf. Learning Representations, 2022. (查阅网上资料, 未找到对应的出版地及页码信息, 请确认补充)
    [14]
    B. Ibarz, J. Leike, T. Pohlen, G. Irving, S. Legg, and D. Amodei, “Reward learning from human preferences and demonstrations in Atari,” in Proc. 32nd Int. Conf. Neural Information Processing Systems, Montréal, Canada, 2018, pp. 8022−8034.
    [15]
    P. Barnett, R. Freedman, J. Svegliato, and S. Russell, “Active reward learning from multiple teachers,” in Proc. Workshop on Artificial Intelligence Safety 2023 co-located with the 37th AAAI Conf. Artificial Intelligence, Washington DC, USA, 2023.(查阅网上资料, 未找到对应的页码信息, 请确认补充)
    [16]
    A. Glaese, N. McAleese, M. Trębacz, J. Aslanides, V. Firoiu, T. Ewalds, et al., “Improving alignment of dialogue agents via targeted human judgements,” arXiv preprint arXiv: 2209.14375, 2022.
    [17]
    T. Everitt, V. Krakovna, L. Orseau, and S. Legg, “Reinforcement learning with a corrupted reward channel,” in Proc. 26th Int. Joint Conf. Artificial Intelligence, Melbourne, Australia, 2017, pp. 4705−4713.
    [18]
    A. Pan, K. Bhatia, and J. Steinhardt, “The effects of reward misspecification: Mapping and mitigating misaligned models,” in Proc. 10th Int. Conf. Learning Representations, 2022.(查阅网上资料, 未找到对应的出版地及页码信息, 请确认补充)
    [19]
    D. Dai, L. Dong, Y. Hao, Z. Sui, B. Chang, and F. Wei, “Knowledge neurons in pretrained transformers,” in Proc. 60th Annu. Meeting of the Association for Computational Linguistics, Dublin, Ireland, 2022, pp. 8493−8502.
    [20]
    E. Mitchell, C. Lin, A. Bosselut, C. Finn, and C. D. Manning, “Fast model editing at scale,” in Proc. 10th Int. Conf. Learning Representations, 2022.(查阅网上资料, 未找到对应的出版地及页码信息, 请确认补充)
    [21]
    J. Wei, X. Wang, D. Schuurmans, M. Bosma, F. Xia, E. H. Chi, Q. V. Le, and D. Zhou, “Chain-of-thought prompting elicits reasoning in large language models,” in Proc. 36th Int. Conf. Neural Information Processing Systems, New Orleans, LA, USA, 2022, Art. no. 1800.
    [22]
    H. Trivedi, N. Balasubramanian, T. Khot, and A. Sabharwal, “Interleaving retrieval with chain-of-thought reasoning for knowledge-intensive multi-step questions,” in Proc. 61st Annu. Meeting of the Association for Computational Linguistics, Toronto, Canada, 2023, pp. 10014−10037.
    [23]
    N. Varshney, W. Yao, H. Zhang, J. Chen, and D. Yu, “A stitch in time saves nine: Detecting and mitigating hallucinations of LLMs by validating low-confidence generation,” arXiv preprint arXiv: 2307.03987, 2023.
    [24]
    P. Lewis, E. Perez, A. Piktus, F. Petroni, V. Karpukhin, N. Goyal, et al., “Retrieval-augmented generation for knowledge-intensive NLP tasks,” in Proc. 34th Int. Conf. Neural Information Processing Systems, Vancouver, BC, Canada, 2020, Art. no. 793.
    [25]
    C. Yang, Y. Liu, K. Huang, B. Sun, Y. Li, X. Chen, and W. Gui, “Intelligent model library for nonferrous metal industry: Construction method and application,” Strategic Study CAE, vol. 24, no. 4, pp. 188−201, 2022.
    [26]
    Z.-P. Chen, J. Han, C.-H. Yang, and W.-H. Gui, “An exploration of a new paradigm for constructing industrial domain-specific embodied intelligent control large models,” Acta Autom. Sinica, vol. 51, no. 11, pp. 2454−2472, Nov. 2025.
    [27]
    Y. Wang, X. Yue, and W. Chen, “Critique fine-tuning: Learning to critique is more effective than learning to imitate,” arXiv preprint arXiv: 2501.17703, 2025.
    [28]
    E. J. Hu, Y. Shen, P. Wallis, Z. Allen-Zhu, Y. Li, S. Wang, L. Wang, and W. Chen, “LoRA: Low-rank adaptation of large language models,” in Proc. 10th Int. Conf. Learning Representations, 2022. (查阅网上资料, 未找到对应的出版地及页码信息, 请确认补充)
    [29]
    Z. Hu, L. Wang, Y. Lan, W. Xu, E.-P. Lim, L. Bing, X. Xu, S. Poria, and R. Lee, “LLM-adapters: An adapter family for parameter-efficient fine-tuning of large language models,” in Proc. Conf. Empirical Methods in Natural Language Processing, Singapore, Singapore, 2023, pp. 5254−5276.
    [30]
    L. Fang, J. Shi, L. Wu, J. Tan, and J. Wan, “Perspectives and prospects on embodied intelligence-empowered smart manufacturing,” J. Intell. Manuf., 2026, DOI: 10.1007/s10845-025-02763-6.
    [31]
    A. Liu, B. Feng, B. Xue, B. Wang, B. Wu, C. Lu, et al., “DeepSeek-V3 technical report,” arXiv preprint arXiv: 2412.19437, 2024.
    [32]
    S. Wang, M. Hu, Q. Li, M. Safari, and X. Yang, “Capabilities of GPT-5 on multimodal medical reasoning,” arXiv preprint arXiv: 2508.08224, 2025.
    [33]
    G. Comanici, E. Bieber, M. Schaekermann, I. Pasupat, N. Sachdeva, I. Dhillon, et al., “Gemini 2.5: Pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities,” arXiv preprint arXiv: 2507.06261, 2025.
    [34]
    A. Grattafiori, A. Dubey, A. Jauhri, A. Pandey, A. Kadian, A. Al-Dahle, et al., “The llama 3 herd of models,” arXiv preprint arXiv: 2407.21783, 2024.
    [35]
    D. Guo, D. Yang, H. Zhang, J. Song, P. Wang, Q. Zhu, et al, “DeepSeek-R1 incentivizes reasoning in LLMs through reinforcement learning,” Nature, vol. 645, no. 8081, pp. 633−638, Sep. 2025. doi: 10.1038/s41586-025-09422-z
    [36]
    A. Yang, A. Li, B. Yang, B. Zhang, B. Hui, B. Zheng, et al., “Qwen3 technical report,” arXiv preprint arXiv: 2505.09388, 2025.
    • Related Articles
    • Supplements(0)

Catalog

    通讯作者: 陈斌, bchen63@163.com
    • 1. 

      沈阳化工大学材料科学与工程学院 沈阳 110142

    1. 本站搜索
    2. 百度学术搜索
    3. 万方数据库搜索
    4. CNKI搜索

    Figures(10)  / Tables(1)

    Get Citation
    PDF
    XML

    Article Metrics

    Article views (31) PDF downloads(7) Cited by()

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return