•  
  •  
 

Bulletin of Chinese Academy of Sciences (Chinese Version)

Keywords

artificial intelligence (AI), frontiers of sciences, foundation of discipline, enabling for complex systems, fundamental research

Abstract

The development of the artificial intelligence （AI） discipline is effectively expanding the capability boundaries of AI. To explore the foundational theories and underlying core technologies for next-generation of general AI, the Chinese Academy of Sciences (CAS) has formulated 20 fundamental scientific problems in AI. Centered on “disciplinary reconstruction of AI foundational theories (Y+AI)” and “AI-enabled applications of complex systems across multiple domains (AI+X)”, these problems are characterized by multiple complexities including openness, high dimensionality, high dynamism, large scale, and cross-scale properties.

First page

2008

Last Page

2017

Language

Chinese

Publisher

Bulletin of Chinese Academy of Sciences

References

[1] 中国科学院学部人工智能赋能科学研究咨询项目学科组. 人工智能学科体系：内涵、架构与发展建议. 中国科学院院刊, 2026, 41(6): 1079-1088. Discipline Group of Advisory Project on AI-empowered Scientific Research, Academic Divisions of the Chinese Academy of Sciences. Disciplinary system of artificial intelligence: Connotation, architecture, and development suggestions. Bulletin of Chinese Academy of Sciences, 2026, 41(6): 1079-1088. (in Chinese)

[2] Smale S. Mathematical problems for the next century. The Mathematical Intelligencer, 1998, 20(2): 7-15.

[3] Colbrook M J, Antun V, Hansen A C. The difficulty of computing stable and accurate neural networks: On the barriers of deep learning and Smale’s 18th problem. PNAS, 2022, 119(12): e2107151119.

[4] Bronstein M M, Bruna J, Cohen T, et al. Geometric deep learning: Grids, groups, graphs, geodesics, and gauges. (2021-04-27)[2026-10-12]. https://arxiv.org/abs/2104.13478.

[5] Wu X Y, Wang Y F, Jegelka S, et al. On the emergence of position bias in transformers. (2025-02-04)[2026-10-12]. https://arxiv.org/abs/2502.01951.

[6] Ding Q G, Wang X Y, Feng R, et al. ComBodied agents: A new paradigm of human-centric agentic AI. (2026-08-12)[2026-10-12]. https://arxiv.org/abs/2608.10915.

[7] Turing A M. On computable numbers, with an application to the Entscheidungsproblem. Proceedings of the London Mathematical Society, 1937, s2-42(1): 230-265.

[8] Yao A C. Classical physics and the church—Turing thesis. Journal of the ACM, 2003, 50(1): 100-105.

[9] Kolmogorov A N. Three approaches to the quantitative definition of information. International Journal of Computer Mathematics, 1968, 2(1-4): 157-168.

[10] Arunachalam S, De W R. A survey of quantum learning theory. Foundations and Trends in Theoretical Computer Science, 2017, 13(3-4): 217-322.

[11] Deutsch D. Quantum theory, the Church–Turing principle and the universal quantum computer. Proceedings of the Royal Society of London A Mathematical and Physical Sciences, 1985, 400: 97-117.

[12] Bennett C H. Time/space trade-offs for reversible computation. SIAM Journal on Computing, 1989, 18(4): 766-776.

[13] Jiang J Q, Sun X M, Teng S H, et al. Optimal space-depth trade-off of CNOT circuits in quantum logic synthesis. Proceedings of the Fourteenth Annual ACM-SIAM Symposium on Discrete Algorithms. Philadelphia, PA: Society for Industrial and Applied Mathematics, 2020: 213-229.

[14] He L X, Xu Y H, He W H, et al. Network model with internal complexity bridges artificial intelligence and neuroscience. Nature Computational Science, 2024, 4(8): 584-599.

[15] Deary I J, Penke L, Johnson W. The neuroscience of human intelligence differences. Nature Reviews Neuroscience, 2010, 11(3): 201-211.

[16] Kaplan J, McCandlish S, Henighan T, et al. Scaling laws for neural language models. (2020-01-23)[2026-10-12]. https://arxiv.org/abs/2001.08361.

[17] Xu N L. How the brain’s primary processing units compute to give rise to intelligence. Nature Reviews Neuroscience, 2024, 25(5): 288-288.

[18] Farzanegan B, Jagannathan S. Explainable and safety aware deep reinforcement learning-based control of nonlinear discrete-time systems using neural network gradient decomposition. IEEE Transactions on Automation Science and Engineering, 2025, 22: 13556-13568.

[19] Fisac J F, Akametalu A K, Zeilinger M N, et al. A general safety framework for learning-based control in uncertain robotic systems. (2017-05-03)[2026-10-12]. https://arxiv.org/abs/1705.01292.

[20] 张宏达, 李德才, 何玉庆. 基于不完备信息预测的多智能体分布式协同. 信息与控制, 2024, 53(1): 86-97. Zhang H D, Li D C, He Y Q. Multi-agent distributed cooperation based on incomplete information prediction. Information and Control, 2024, 53(1): 86-97. (in Chinese)

[21] 侯忠生, 许建新. 数据驱动控制理论及方法的回顾和展望. 自动化学报, 2009, 35(6): 650-667. Hou Z S, Xu J X. On data-driven control theory: The state of the art and perspective. Acta Automatica Sinica, 2009, 35(6): 650-667. (in Chinese)

[22] Hubinger E, Denison C, Mu J, et al. Sleeper agents: Training deceptive LLMs that persist through safety training. (2024-01-10)[2026-10-12]. https://arxiv.org/abs/2401.05566.

[23] Wolf Y, Wies N, Avnery O, et al. Fundamental limitations of alignment in large language models. (2023-04-19)[2026-10-12]. https://arxiv.org/abs/2304.11082.

[24] Glukhov D, Shumailov I, Gal Y, et al. Position: Fundamental limitations of LLM censorship necessitate new approaches// Proceedings of the 41st International Conference on Machine Learning. 2024: 15767-15787.

[25] Cao F F, Jia F L, He X. Sensor fault detection and diagnosis of linear parabolic PDE systems with unknown inputs. IEEE Transactions on Automatic Control, 2024, 69(2): 1014-1021.

[26] Johnson S G B, Karimi A H, Bengio Y, et al. Imagining and building wise machines: The centrality of AI metacognition. Trends in Cognitive Sciences, 2026, 30(10): 854-868.

[27] Gu R L, Zhou Y Q, Yao S C, et al. The human-AI empathy loop: A path toward stronger human-AI relationships. Science Bulletin, 2026, 71(4): 683-687.

[28] Ten A, Kaushik P, Oudeyer P Y, et al. Humans monitor learning progress in curiosity-driven exploration. Nature Communications, 2021, 12: 5972.

[29] 张钹, 朱军, 苏航. 迈向第三代人工智能. 中国科学: 信息科学, 2020, 50(9): 1281-1302. Zhang B, Zhu J, Su H. Toward the third generation of artificial intelligence. Scientia Sinica (Informationis), 2020, 50(9): 1281-1302. (in Chinese)

[30] Wang H C, Fu T F, Du Y Q, et al. Scientific discovery in the age of artificial intelligence. Nature, 2023, 620: 47-60.

[31] Agarwal D, Majumder B P, Adamson R, et al. AutoDiscovery: Open-ended scientific discovery via Bayesian surprise. Advances in Neural Information Processing Systems, 2025: 28504-28542.

[32] Lu C, Lu C, Lange R T, et al. Towards end-to-end automation of AI research. Nature, 2026, 651: 914-919.

[33] Agrawal A, Nadeem M, Al Nuaim A, et al. Artificial intelligence driven multi agent framework for adaptive cyber attack simulation and automated incident response in cyber range environments. Scientific Reports, 2026, 16: 11673.

[34] Hagendorff T, Derner E, Oliver N. Large reasoning models are autonomous jailbreak agents. Nature Communications, 2026, 17: 1435.

[35] Betley J, Warncke N, Sztyber-Betley A, et al. Training large language models on narrow tasks can lead to broad misalignment. Nature, 2026, 649: 584-589.

[36] Mahto S K, Singh B, Kushwaha N, Kumar R. AI-powered cybersecurity for next-generation threats// Dash S P, Mishra V, Shaw B, et al. Digital Immune System: Principles and Practices. Hoboken: John Wiley & Sons, 2025: 289-304.

[37] Wang Z, Shi T N, He J X, et al. CyberGym: Evaluating AI agents’ real-world cybersecurity capabilities at scale. (2025-06-03)[2026-10-12]. https://arxiv.org/abs/2506.02548.

[38] Runge J, Bathiany S, Bollt E, et al. Inferring causation from time series in Earth system sciences. Nature Communications, 2019, 10: 2553.

[39] Karniadakis G E, Kevrekidis I G, Lu L, et al. Physics-informed machine learning. Nature Reviews Physics, 2021, 3(6): 422-440.

[40] Raissi M, Perdikaris P, Karniadakis G E. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 2019, 378: 686-707.

[41] 徐萍, 许丽, 杨若南, 等. 新药研发发展态势分析. 中国科学院院刊, 2024, 39(5): 821-831. Xu P, Xu L, Yang R N, et al. Analysis of research and development trends of new drugs. Bulletin of Chinese Academy of Sciences, 2024, 39(5): 821-831. (in Chinese)

[42] Abramson J, Adler J, Dunger J, et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature, 2024, 630: 493-500.

[43] Hao M S, Gong J, Zeng X, et al. Large-scale foundation model on single-cell transcriptomics. Nature Methods, 2024, 21(8): 1481-1491.

[44] Zhang K, Yang X, Wang Y F, et al. Artificial intelligence in drug development. Nature Medicine, 2025, 31(1): 45-59.

[45] 朱新广, 鲍一明, 金哲侬, 等. 人工智能赋能生命科学研究范式变革、应用场景及发展布局. 中国科学院院刊, 2026, 41(6): 1127-1138. Zhu X G, Bao Y M, Jin Z N, et al. Artificial intelligence empowered biological research: Paradigm shifts, application scenarios, and strategic layout. Bulletin of Chinese Academy of Sciences, 2026, 41(6): 1127-1138. (in Chinese)

[46] Xing X D, Shi F D, Huang J H, et al. On the caveats of AI autophagy. Nature Machine Intelligence, 2025, 7(2): 172-180.

[47] Guo L X, Chen H Y, Chen Y, et al. A neural network for modeling human concept formation, understanding and communication. Nature Computational Science, 2026, 6(5): 497-511.

[48] Zomer N, De Domenico M. Unraveling the emergence of collective behavior in networks of cognitive agents. npj Artificial Intelligence, 2026, 2: 36.

[49] Brinkmann L, Eisenmann T F, Nussberger A M, et al. Propagation and preservation of AI-discovered problem-solving strategies in human culture. Nature Communications, 2026, 17: 8255.

Download
Request a Copy

Share

COinS