[Paper Review] Intelligent Computing: The Latest Advances, Challenges and Future

Shiqiang Zhu, Ting Yu+19 more|arXiv (Cornell University)|Nov 21, 2022

Advanced Memory and Neural ComputingEngineering321 references19 citations

TL;DR

This paper presents a comprehensive survey of intelligent computing, unifying theory, systems, and applications across data, perceptual, cognitive, and autonomous intelligence. It identifies key challenges in interpretability, scalability, and system complexity, and proposes a roadmap for future research in intelligent, scalable, and trustworthy computing systems.

ABSTRACT

Computing is a critical driving force in the development of human civilization. In recent years, we have witnessed the emergence of intelligent computing, a new computing paradigm that is reshaping traditional computing and promoting digital revolution in the era of big data, artificial intelligence and internet-of-things with new computing theories, architectures, methods, systems, and applications. Intelligent computing has greatly broadened the scope of computing, extending it from traditional computing on data to increasingly diverse computing paradigms such as perceptual intelligence, cognitive intelligence, autonomous intelligence, and human-computer fusion intelligence. Intelligence and computing have undergone paths of different evolution and development for a long time but have become increasingly intertwined in recent years: intelligent computing is not only intelligence-oriented but also intelligence-driven. Such cross-fertilization has prompted the emergence and rapid advancement of intelligent computing. Intelligent computing is still in its infancy and an abundance of innovations in the theories, systems, and applications of intelligent computing are expected to occur soon. We present the first comprehensive survey of literature on intelligent computing, covering its theory fundamentals, the technological fusion of intelligence and computing, important applications, challenges, and future perspectives. We believe that this survey is highly timely and will provide a comprehensive reference and cast valuable insights into intelligent computing for academic and industrial researchers and practitioners.

Motivation & Objective

• To provide a holistic overview of intelligent computing as a transformative paradigm integrating intelligence and computing.
• To identify core challenges in interpretability, generality, and system complexity across intelligent computing systems.
• To analyze the technological fusion of AI, big data, IoT, and advanced computing architectures.
• To map the evolution from data-centric computing to autonomous and human-computer fusion intelligence.
• To outline future research directions for scalable, reliable, and explainable intelligent computing systems.

Proposed method

• Conducting a systematic literature review across theory, systems, and applications of intelligent computing.
• Categorizing intelligent computing into perceptual, cognitive, autonomous, and human-computer fusion intelligence paradigms.
• Analyzing horizontal and vertical computing architectures for large-scale system scalability and resource optimization.
• Proposing a multi-level abstraction framework to model micro-macro linkages in complex systems.
• Introducing a human-guided task decomposition method to manage large-scale, non-linear, and interdependent computational problems.
• Establishing a higher-order theoretical perspective to analyze system interpretability and uncertainty in large-scale systems.

Experimental results

Research questions

• RQ1 How can intelligent computing systems achieve higher levels of autonomy, interpretability, and generalization beyond narrow AI?
• RQ2 What are the fundamental architectural and theoretical challenges in scaling intelligent computing systems to handle real-world complexity?
• RQ3 How can uncertainty and non-linear interactions in multi-level subsystems be modeled and managed effectively?
• RQ4 What role can human-in-the-loop guidance play in decomposing and solving large-scale, interdependent computational tasks?
• RQ5 How can a unified theoretical framework be developed to ensure interpretability and reliability in large-scale intelligent systems?

Key findings

• Intelligent computing has evolved from data-centric computing to encompass perceptual, cognitive, and autonomous intelligence, enabling broader societal impact.
• Horizontal computing architectures, such as Folding@home, have demonstrated exaflop-scale performance by leveraging distributed volunteer computing.
• Current deep learning models face significant limitations in interpretability, generalization, and autonomy, falling short of human-level intelligence.
• Large-scale systems suffer from high complexity due to non-linear subsystem interactions and uncertainty, which current optimization methods fail to capture.
• There is no solid theoretical foundation for simplifying large-scale optimization problems through decomposition, especially under multiple uncertainties.
• A new higher-order theoretical perspective is needed to analyze system interpretability and structural evolution in complex, real-world intelligent systems.