徐明坤
博士，青年研究員，研究組組長
Email: xumingkun@@gdiist.cn
個人簡介：
徐明坤，青年研究員，類腦智能算法與模型研究組組長。2018年本科畢業于西安電子科技大學，2023年于清華大學獲得博士學位。2023年—2025年期間在廣東省智能科學與技術研究院從事博士后研究。徐明坤博士專注于類腦算法、人工智能模型與BI for Science智能應用，相關研究工作在Nature Communications、Science Robotics、IEEE Transactions on Neural Networks and Learning Systems等高水平期刊和ICML、AAAI、CVPR、IJCAI、EMNLP等人工智能頂會上發表論文30余篇，并已獲得多項國家發明專利授權。
類腦智能算法與模型研究組
本研究組專注于探索“人工智能 × 大模型 × 腦科學”的交叉前沿，致力于構建兼具生物可解釋性與通用智能能力的新一代人工智能體系。以神經科學發現為指導，以機器學習與高性能計算為手段，從神經元尺度到系統級智能橫向貫通，自底向上打造具備推理、記憶、持續學習與自主適應的類腦智能框架。通過跨學科協同，力求揭示類腦計算的核心機理，并將其轉化為服務于機器人、自動駕駛、生命健康等領域的關鍵技術與創新應用。
當前，研究組的核心研究方向包括：
1. 類腦智能算法
依托脈沖神經網絡的時空特性，面向記憶鞏固、連續學習、樹突非線性整合、穩態機制、Binding 機制與認知地圖等關鍵腦啟發原理，設計低能耗、可擴展的學習規則與計算模型。研究重點涵蓋突觸可塑性、表征機制、元學習驅動的任務快速遷移、以及動態結構重塑等機制，力求在跨模態、多任務環境中構建具備“可解釋-可遷移-可增量演化”三維能力的類腦智能體，為智能應用與大模型融合奠定生物學與工程并重的算法基礎。
2. 深度學習與大模型
重點圍繞基礎模型架構設計、參數微調與快速推理三條主線展開。架構層面，以 Transformer／Mamba 等代表網絡為原型，探索從視覺到跨模態的創新高效架構及統一大模型框架；在訓練層面，結合 LoRA 等輕量策略與檢索增強方法，實現低成本微調與增量學習；在推理層面，聚焦 KV-Cache 優化、低比特量化等技術，構建低時延、低功耗的端-云協同推理體系。此外，將探索生成模型、圖表示學習、強化學習驅動的 Agent 架構等相關技術，賦予模型感知-記憶-決策的閉環能力，為類腦應用提供堅實算法底座。
3. 類腦智能應用與BI for Science交叉賦能
將前述算法模型應用到自動駕駛、機器人協作與終端設備在線學習等復雜場景，打造從感知—決策—執行的全鏈路解決方案，顯著提升智能體在動態不確定環境中的自主適應與能效比。此外，融合類腦算法與大模型技術，面向醫療診斷、核酸分子檢測、蛋白質屬性預測等問題，構建任務驅動的智能分析工具，為生命健康、材料研發等領域提供突破性方案。
代表論著：
1. Xu M, Liu F, Hu Y, Li H, Wei Y, Zhong S, Pei J* and Deng L*. Adaptive Synaptic Scaling in Spiking Networks for Continual Learning and Enhanced Robustness[J]. IEEE Transactions on Neural Networks and Learning Systems, 2024. (中科院一區SCI，TOP期刊)
2. Xu M, Wu Y, Deng L, Liu F, Li G and Pei J*. Exploiting spiking dynamics with spatial-temporal feature normalization in graph learning. In Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence (IJCAI-21), pages 3207-3213, 2021. (人工智能國際會議，CCF A類)
3. Yang Y#, Xu M#（共同一作）, Jia S, Wang B, Xu L, Wang X, Liu H, Liu Y, Guo Y, Wang L, Duan S, Liu K, Zhu M, Pei J, Duan W, Liu D, Li H*. A new opportunity for the emerging tellurium semiconductor: making resistive switching devices[J]. Nature Communications, 2021, 12(1): 1-12. (中科院一區SCI，TOP期刊)
4. Wu Y#, Zhao R#, Zhu J#, Chen F#, Xu M#（共同一作）, Li G, Song S, Deng L, Wang G, Zheng H, Ma S, Pei J, Zhang Y, Zhao M, Shi L*. Brain-inspired global-local learning incorporated with neuromorphic computing[J]. Nature Communications, 2022, 13(1): 1-14. (中科院一區SCI， TOP期刊)
5. Xu M*. Exploiting homeostatic synaptic modulation in spiking neural networks for semi-supervised graph learning[C]. Proceedings of the 32nd ACM International Conference on Information and Knowledge Management(CIKM). 5193-5195. (2023) (人工智能國際會議，CCF B類)
6. Chen Y, Song A, Yin H, Zhong S, Chen F, Xu Q, Wang S*, Xu M*（通訊作者）. Multi-view incremental learning with structured hebbian plasticity for enhanced fusion efficiency[C]//Proceedings of the AAAI Conference on Artificial Intelligence. 2025, 39(2): 1265-1273. (人工智能國際會議，CCF A類)
7. Wang Y, Fang X, Yin H, Li D, Li G, Xu Q*, Xu Y, Zhong S, Xu M*（通訊作者）. BIG-FUSION: Brain-Inspired Global-Local Context Fusion Framework for Multimodal Emotion Recognition in Conversations[C]//Proceedings of the AAAI Conference on Artificial Intelligence. 2025, 39(2): 1574-1582. (人工智能國際會議，CCF A類)
8. Wang Y, Tan S, Shen J, Xu Y, Song H, Xu Q, Tiwari P, Xu M*（通訊作者）. Enhancing Graph Contrastive Learning for Protein Graphs from Perspective of Invariance[C]// Forty-Second International Conference on Machine Learning (ICML). (Conference Track). 2025. (人工智能國際會議，CCF A類)
9. Xu M, Zheng H, Pei J, Deng L*. A Unified Structured Framework for AGI: Bridging Cognition and Neuromorphic Computing [C]//Artificial General Intelligence: 16th International Conference, AGI 2023, Stockholm, Sweden, June 16–19, 2023, Proceedings. Cham: Springer Nature Switzerland, 2023: 345-356.
10. Wang Y, Li D, Shen J, Xu Y, Zhong S, Xu M*（通訊作者）. ClingTP: Curriculum Learning based Multi-style Title Prefix Generation[C]//ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2025: 1-5. (人工智能國際會議，CCF B類)
11. Xu M, Liu F, Pei J*. Endowing spiking neural networks with homeostatic adaptivity for APS-DVS bimodal scenarios. [C]//Companion Publication of the 2022 International Conference on Multimodal Interaction(ICMI). 2022: 12-17.
12. Ran X#, Xu M#（共同一作）, Mei L, Xu Q, Liu Q*. Detecting out-of-distribution samples via variational auto-encoder with reliable uncertainty estimation[J]. Neural Networks, 2022, 145: 199-208.
13. Liu F#, Xu M#（共同一作）, Li G, Pei J, Shi L, Zhao R*. Adversarial symmetric GANs: Bridging adversarial samples and adversarial networks[J]. Neural Networks, 2021, 133: 148-156.
14. Wang Y#, Zhang Z#, Xu M#（共同一作）, Yang Y, Ma M, Li H, Pei J, Shi L*. Self-doping memristors with equivalently synaptic ion dynamics for neuromorphic computing[J]. ACS applied materials & interfaces, 2019, 11(27): 24230-24240. (中科院一區SCI，TOP期刊，封面文章)
15. Song A, Chen Y, Wang Y, Zhong S, Xu M*(通訊作者). Orchestrating Plasticity and Stability: A Continual Knowledge Graph Embedding Framework with Bio-Inspired Dual-Mask Mechanism[C]//The 16th Asian Conference on Machine Learning (Conference Track). 2024.
16. Zhong S*, Su L, Xu M, Loke D, Yu B, Zhang Y, Zhao R*. Recent Advances in Artificial Sensory Neurons: Biological Fundamentals, Devices, Applications, and Challenges[J]. Nano-Micro Letters, 2025, 17(1): 61. (中科院一區SCI，TOP期刊)
17. Liu Z, Chen J, Xu M, Ho S, Wei Y*, Ho HP, Yong KT*. Engineered multi-domain lipid nanoparticles for targeted delivery[J]. Chemical Society Reviews, 2025. (中科院一區SCI，TOP期刊)
18. Yu F, Wu Y, Ma S, Xu M, Li H, Qu H, Song C, Wang T, Zhao R, Shi L*. NeuroGPR: Brain-inspired General Place Recognition with Neuromorphic Computing [J]. Science Robotics, 2023, 8(78): eabm6996. (中科院一區SCI，TOP期刊)
19. Zhao R, Yang Z, Zheng H, Wu Y, Liu F, Wu Z, Li L, Chen F, Song S, Zhu J, Zhang W, Huang H, Xu M, Sheng K, Yin Q, Pei J, Li G, Zhang Y, Zhao M, Shi L*. A framework for the general design and computation of hybrid neural networks[J]. Nature Communications, 2022, 13(1): 1-12. (中科院一區SCI，TOP期刊)
