What Is Computational Artificial Intelligence?

Computational artificial intelligence (CI) is typically defined as the ability of a computer to undertake specific tasks through learning or experimental observation, rather than explicit instruction.  CI is also expressed as the study of adaptive mechanisms to enhance or facilitate intelligent behavior in complex and fluctuating environments. 

Generally, Artificial intelligence (AI) seeks to enable a computer or any computer-controlled machine to execute tasks that would otherwise require some aspect of human-like intelligence. Computational intelligence depicts paradigms exhibited in AI that facilitate the ability to learn or adapt to new situations, generalize, discover, abstract, and associate. Generally, CI entails a combination of nature-inspired computational methodologies and approaches to resolve complex real-world complications that cannot be addressed using mathematical or traditional modeling approaches. 

Methods used in computational intelligence solve problems related to human reasoning, i.e., utilization of imprecise or inadequate knowledge. However, CI prevails by adaptively producing control actions. Computational intelligence utilizes the following complimentary paradigms: artificial neural networks, fuzzy systems, and evolutionary computing. 

Paradigms of Computational Artificial Intelligence

Artificial Neural Networks

Biological neural networks inspire artificial neural networks (NNs). Computational intelligence experts design NNs using the three main components of biological neural networks: a cell body for information processing, the axon-for enabling relay of signal, and the synapse-for signal control. Artificial neural networks thus distribute information processing systems to facilitate the process and learning from experiential data. Areas of research include data analysis NNs, classification NNs, associative memory NNs, recurrent NNs, control NNs, etc.  

Fuzzy Systems

Fuzzy systems (FS) use human language as the source of inspiration based on the observation that humans make judgments out of imprecise and non-numerical information. Imprecise information or vagueness is represented using mathematical means called fuzzy models or sets; hence, the term fuzzy. Fuzzy models can represent, recognize, manipulate, and utilize vague and uncertain data and information. Areas of research include linguistic summarization, fuzzy controllers, fuzzy sets and systems and fuzzy neural networks. 

Evolutionary Computation

Evolutionary computation (EC) draws its inspiration from the biological concept of natural evolution to guide new artificial evolutionary methodologies. EC entails programs such as evolution strategy and evolutionary algorithms that provide solutions to problems that traditional mathematical techniques cannot address. Other areas of EC include multi-objective optimization, genetic algorithms, swarm intelligence, genetic evolution, differential evolution, and evolutionary programming.

Applications

Computational artificial intelligence is applied in:

• To develop intelligent systems in multiple areas, including gaming and data processing systems.
• Bio-medicine areas such as computational genomics, prediction of gene expressions, investigating protein structures and modeling neural systems.
• Engineering to identify material constants using NNs, equipment fault diagnosis using FS, and digital circuit synthesis using EC.