Data Science Lab

Introduction

We are in the era of complex systems and complexity and intelligence are two fundamental and omnipresent ingredients forming our everyday, social, business, and cyber worlds [1-10]. Some increasingly critical research issues facing researchers in different disciplines include:

how to understand their characteristics and complexities,
how to formalize and quantify such complex systems, and
how to `compute and engineer’ their corresponding problem-solving systems.

Faced with open complex giant systems like the Internet [11], we often do not know what we do not know [12,41] and there are many unknowns in complex data, behaviours, systems and problem-solvers [39,12,41]. Many critical research questions are to explore, exploit such unknowns.

In our scientific history, three main scientific metaphors, thinking and methodological paradigms have been proposed to understand and engineer complex systems [1-6,39,40]:

reductionism: dedicating priority to the parts of a system rather than the whole;
holism: focusing on systems thinking and an emphasis on understanding complexities in the whole rather than merely in part; and
systematology: combining top-down holistic methodologies with bottom-up reductionistic approaches to consider the complexities in the whole system and its parts, as well as their respective connections.

Systematology and Qualitative-to-Quantitative Metasynthesis

Systematology [1-6] appears to be really useful for addressing the complexities of open complex giant systems[1-2], in which it may not be effective to only apply reductionism or holism. A qualitative-to-quantitative metasynthesis [1-6] which explores the synergy of human intelligence and machine intelligence in a human-centered, human-machine-cooperative way may be helpful for understanding system complexities and building problem-solving systems, as trialed in building a hall for workshop of metasynthetic engineering (HWME) for macroeconomic decision-support [14-17] (and many other references unlisted here).

The following diagram shows the problem-solving process of qualitative-to-quantitative metasynthesis for understanding and engineering open complex giant systems (OCGS).

Figure 1. A problem-solving process of qualitative-to-quantitative metasynthesis [13,39].

Metasynthetic AI, Computing and Engineering

Metasynthetic AI, computing and engineering [13,39] take human-centered, human-machine-cooperative, and qualitative-to-quantitative metasynthesis as the main methodology and process to understand, compute and engineer such complex systems and their intelligences and problem-solving systems. From the perspectives of AI, computing and engineering, some important questions are worth studying, including

how to undertake the system analysis and system design of complex problem-solving systems,
how to utilize existing system/software engineering methods, computing paradigms, X-AI and X-intelligences, and diverse techniques and tools to implement such systems,
how to evaluate and assure the quality and impact of such systems.

Answering such questions may be inspired by the theories of systematology and qualitative-to-quantitative metasynthesis [1-6]. Some of relevant explorations and applications of such methodologies and engineering methods including understanding the diverse, hierarchical, evolving, and interconnected properties [11,39,40], characteristics, complexities, intelligences [12,37-38,41], interactions, and couplings [18] in the underlying entities (e.g., data, objects) and their behaviours and dynamics, as well as their composite subsystems and systems [13,39].

In the process of computing and engineering complex intelligent systems, we aim to understand, formalize and compute the metasynthetic interactions (m-interactions) [13,39] and couplings [18], x-complexities [12,41,42], and x-intelligences [12,41,42] in complex systems, undertake metasynthetic computing (m-computing) [13,39], and build the problem-solving metasynthetic spaces (m-space) [13,39,40].

Some of Explorations and Applications

Over a decade, we have been exploring complex data, behaviours and systems and their understanding and computing from the perspective of complex systems, including expanding and applying the methodologies of systematology and qualitative-to-quantitative metasynthesis, and undertaking metasynthetic computing and engineering in various domains and research areas. Examples are:

Data science as complex systems: to understand the thinking and mindset for data science, i.e., data science thinking [41], as a discipline, and understand the critical challenges, foundations, and opportunities of data science from cross-domain, inter-disciplinary, and human-data-intelligence-cooperative perspectives [36-38,41]
Domain-driven, actionable knowledge discovery: to address the limitations in conventional data mining and machine learning, we advocate the incorporation, formalization, analysis, learning, and metasynthesis of x-complexities, x-intelligences, x-analytics, and various evaluation metrics from both statistical and business impact, and both subjective and objective perspectives [21-25,41,42];
Behavior informatics and computing: to understand, formalize, compute, analyze, learn and manage complex behaviors of individuals, organizations, systems and societies, occurring and nonoccurring behaviors, behavioral impact and utility, and complex behavioral and social systems [26-34,43];
Coupling and interaction learning: to formalize, represent, analyze, learn and manage complex interactions and coupling relationships in complex data, behaviors, and systems and transform existing theories and systems of pattern recognition, data mining, machine learning, risk management, outlier detection, and recommendation, etc.; [18] (more about coupling learning can be found in the topic on coupling and interaction learning);
New-generation recommender systems: to understand the complexities, intelligences surrounding the entities, interactions, couplings of recommender systems, including its users, products, context, and dynamics for actionable recommendation [35].

Other Research Topics

There are many interesting research topics in metasynthetic AI, computing and engineering of open complex giant systems such as open complex intelligent systems, including but are not limited to the following areas:

Enabling systematology: synthesizing reductionism and holism, conducting qualitative-to-quantitative metasynthesis;
Cognitive modeling and social cognitive intelligence: individual thinking, group thinking, social thinking, creative thinking, critical thinking, imaginary thinking, emergence of social intelligence;
X-complexities: such as openness, hierarchy, sociality, interaction, heterogeneity, scale, evolution, human-machine interaction and cooperation, and their metasynthesis;
X-intelligences: such as behavior intelligence, data intelligence, domain intelligence, symbolic intelligence, situated intelligence, connectionist intelligence, nature-inspired intelligence, social intelligence, organizational intelligence, network intelligence, human intelligence, and their metasynthesis;
Intelligence metasynthesis: developing methods and mechanisms to synthesize diverse intelligences including human intelligence, domain intelligence, data intelligence, behavior intelligence, machine intelligence, network intelligence, social intelligence, and organizational intelligence; building consensus, managing divergence and convergence between intelligences;
X-computing paradigms: such as agent-based computing, service-oriented computing, behavior computing, organizational computing, social computing, cloud computing, metasynthetic computing, and their integration;
X-analytics: general methodologies, analytical systems and processes for risk analytics, behavior analytics, social analytics, business analytics, corporate analytics, government analytics, financial analytics, etc. and various analytical techniques such as text analytics, tabular analytics, multimedia analytics, behavior analytics, etc. and their metasynthesis;
Universal representation and learning: unifying representation, learning, inference, reasoning, optimization and refinement in complex analytical and learning systems and processes;
Knowledge fusion and new knowledge production: fusing human knowledge, experience and expertise with machine, organizational and social information and capabilities to discover, infer, produce new knowledge and intelligence;
Actionable knowledge and intelligence: discovering and delivering knowledge and intelligence that can fit real-world systems, problem-solving processes and systems, and deliver significant benefits and impacts;
System abstraction, representation, modeling and implementation: visual and formal methods, organizational abstraction, interaction modeling, emergence mechanisms, etc. for complex systems;
Metasynthetic engineering: building a hall for workshop of metasynthetic engineering (HWME) or intelligent problem-solving systems to ‘compute and engineer’ open complex giant systems.

References

[1]. Qian, X., Yu, J., Dai, R.: A new scientific field–open complex giant systems and the methodology (in Chinese). Chin. J. Nat. 13(1), 3–10 (1990)
[2]. Qian, X.: Revisiting issues on open complex giant systems (in Chinese). Pattern. Recogn. Artif. Intell. 4(1), 5–8 (1991)
[3]. Qian, X.: Building Systematology (in Chinese). Shanghai Jiaotong University Press, Taiyuan, China (2007)
[4]. Dai, R., et al.: Metasynthesis of Intelligent Systems. Zhejiang Science & Technology Press (in Chinese), Hangzhou, China (1995)
[5]. Dai, R., Li., Y. and Li, Q.: Social Intelligence and Metasynthetic System (in Chinese). Post & Telecom, Beijing, China (2013)
[6]. Dai, R.: Qualitative-to-quantitative metasynthetic engineering (in Chinese). Pattern. Recogn. Artif. Intell. 6(2), 60–65 (1993)
[7]. Mitchell, M.: Complexity: A Guided Tour. Oxford University Press, New York (2011)
Miller, J.H., Page, S.E.: Complex Adaptive Systems: An Introduction to Computational Models of Social Life. Princeton University Press, Princeton (2007)
[8]. Holland, J.H.: Signals and Boundaries: Building Blocks for Complex Adaptive Systems. The MIT Press, Cambridge (2014)
Waldrop, M.M.: Complexity: The Emerging Science at the Edge of Order and Chaos. Simon & Schuster, New York (1992)
[9]. Meadows, D.H.: Thinking in Systems: A Primer. Chelsea Green Publishing, White River Junction (2008)
[10]. Page, S.E.: Diversity and Complexity (Primers in Complex Systems). Princeton University Press, Princeton (2010)
[11] Dai, R. and Cao, L.: “Internet—-An Open Complex Giant System”, Science in China (Series E), Sciences In China Series E, 33(4), 289-296, 2003 (in Chinese)
[12] Longbing Cao. Data Science: Challenges and Directions. Communications of the ACM, Vol. 60 No. 8, Pages 59-68. BibTeX
[13]. Longbing Cao, Ruwei Dai, Mengchu Zhou. Metasynthesis: M-Space, M-Interaction and M-Computing for Open Complex Giant Systems, IEEE Trans. On Systems, Man, and Cybernetics–Part A, 39(5): 1007 – 1021, 2009. BibTeX
[14]. Longbing Cao, Ruwei Dai. Agent-Oriented Metasynthetic Engineering for Decision Making, International Journal of Information Technology and Decision Making, 2(2):197-215, World Scientific Publishing, 2003. BibTeX
[15]. Longbing Cao, Ruwei Dai. Human-Computer Cooperated Intelligent Information System Based on Multi-Agents, ACTA AUTOMATICA SINICA, 29(1):86-94, 2003, China (in English)
[16]. Longbing Cao, Ruwei Dai. Agent-Oriented Approach for Dealing with Open Giant Intelligent Systems, Journal of Pattern Recognition and Artificial Intelligence, 15(3): 75-81,2002 (in Chinese)
[17]. Longbing Cao, Ruwei Dai. Software Architecture of the Hall for Workshop of Metasynthetic Engineering, Journal of Software, China, 13(8):1430-1435, 2002 (in Chinese)
[18] Longbing Cao. Coupling Learning of Complex Interactions, Journal of Information Processing and Management, 51(2): 167-186 (2015). BibTeX
[19]. Longbing Cao, Chengqi Zhang, Ruwei Dai. Intelligence Metasynthesis in Building Business Intelligence Systems, Web Intelligence Meets Brain Informatics, LNCS4845, 454-470, 2007. BibTeX
[20]. Longbing Cao, Chengqi Zhang, Ruwei Dai. Organization-Oriented Analysis of Open Complex Agent Systems. Int. J. on Intelligent Control and Systems, Vol.10 No.2, pp114-122, 2005.
[21] Longbing Cao. Domain-driven data mining: challenges and prospects, IEEE Trans. on Knowledge and Data Engineering, 22(6): 755-769, 2010. BibTeX
[22] Longbing Cao et al. Domain-Driven, Actionable Knowledge Discovery, IEEE Intelligent Systems, 22 (4):78-89, 2007. BibTeX
[23] Longbing Cao, Chengqi Zhang. The evolution of KDD: Towards domain-driven data mining. International Journal of Pattern Recognition and Artificial Intelligence, 21(4): 677-692, 2007. BibTeX
[24] Longbing Cao, Chengqi Zhang. Domain-Driven Data Mining: A Practical Methodology, International Journal of Data Warehousing and Mining, 2(4): 49-65, 2006. BibTeX
[25] Longbing Cao, Chengqi Zhang. Domain-driven actionable knowledge discovery in the real world, PAKDD2006, LNAI 3918, 821 – 830. BibTeX
[26] Longbing Cao. In-depth Behavior Understanding and Use: the Behavior Informatics Approach, Information Science, 180(17); 3067-3085, 2010. BibTeX
[27] Longbing Cao. Behavior Informatics to Discover Behavior Insight for Active and Tailored Client Management. KDD2017, 2017. BibTeX
[28] Longbing Cao, et al. Behavior Informatics: A New Perspective. IEEE Intelligent Systems (Trends and Controversies), 29(4): 62-80, 2014. BibTeX
[29] Longbing Cao, Philip S. Yu, Vipin Kumar. Nonoccurring Behavior Analytics: A New Area. IEEE Intelligent Systems 30(6): 4-11 (2015). BibTeX
[30] Longbing Cao. Health and Medical Behavior Informatics (in Biomedical Information Technology, 2nd Edition, David Dagan Feng (Ed.)), 735-761, Elsevier, 2020. BibTeX
[31] Longbing Cao. Zhao Y., Zhang, C. Mining Impact-Targeted Activity Patterns in Imbalanced Data, IEEE Trans. on Knowledge and Data Engineering, 20(8): 1053-1066, 2008. BibTeX
[32] Junfu Yin, Zhigang Zheng, Longbing Cao. USpan: An Efficient Algorithm for Mining High Utility Sequential Patterns, KDD 2012, 660-668. BibTeX
[33] Longbing Cao. Non-IIDness Learning in Behavioral and Social Data, The Computer Journal, 57(9): 1358-1370 (2014). BibTeX
[34] Can Wang, and Longbing Cao.Modeling and Analysis of Social Activity Process, in Longbing Cao and Philip S Yu (eds) Behavior Computing, 21-35, Springer, 2012
[35] Longbing Cao. Non-IID Recommender Systems: A Review and Framework of Recommendation Paradigm Shifting. Engineering, 2: 212-224, doi:10.1016/J.ENG.2016.02.013., 2016. BibTeX
[36] Longbing Cao. Data Science: A Comprehensive Overview. ACM Computing Surveys, 50(3), 43:1-42, 2017. BibTeX
[37] Longbing Cao. Data Science: Profession and Education. IEEE Intelligent Systems, 34(5): 35-44, 2019. BibTeX
[38] Longbing Cao. Data Science: Nature and Pitfalls. IEEE Intelligent Systems, Volume: 31, Issue: 5, 66-75, 2016. BibTeX

	[39] Longbing Cao. Metasynthetic Computing and Engineering of Complex Systems, ISBN: 978-1-4471-6550-7, Springer, 2015. BibTeX Chapter 14 illustrates the potential application of metasynthetic computing and engineering in actionable knowledge discovery in complex data, behaviours and systems.
	[40] Longbing Cao, Ruwei Dai. Open Complex Intelligent Systems: Fundamentals, Concepts, Analysis, Design and Implementation (in Chinese), ISBN: 9787115172488, Posts & Telecom Press, China, 2008. BibTeX
	[41] Longbing Cao. Data Science Thinking: The Next Scientific, Technological and Economic Revolution, ISBN: 978-3-319-95092-1, Springer International Publishing, 2018. Download the frontmatter and backmatter. BibTeX Section 5.3 discusses about methodologies including reductionism, holism and systematism for data science and synthesizing X-intelligence for complex data science systems.
	[42] Longbing Cao, Philip S Yu, Chengqi Zhang and Yanchang Zhao. Domain Driven Data Mining, ISBN: 978-1-4419-5737-5, Springer, 2010. BibTeX
	[43] Longbing Cao, Philip S Yu (Eds). Behavior Computing: Modeling, Analysis, Mining and Decision, ISBN: 978-1-4471-2968-4, Springer, 2012. BibTeX