What Does Computational Intelligence Mean?
Computational intelligence (CI) is a subfield of artificial intelligence (AI) that focuses on developing biology-inspired AI systems that are able to mimic the way entities in the natural world solve complex problems.
CI strategies are often used in situations where rule-based or statistical approaches to machine learning (ML) are not effective because the problems at hand can’t be addressed through traditional linear algorithms.
Artificial Intelligence (AI) | Computational Intelligence (CI) | |
Definition | A broad field of computer science that focuses on how technology can be used to augment human intelligence and complete tasks on behalf of human beings. | A subset of artificial intelligence that focuses on researching how humans learn and the way bees and ants – and other social entities in the natural world – communicate and cooperate to solve complex problems. |
Scope | Encompasses a wide range of techniques and strategies, including decision trees and recommendation engines. | Primarily focuses on specific types of techniques and strategies within AI, such as fuzzy logic and swarm intelligence. |
Learning Approach | Emphasizes rule-based solutions and linear machine learning algorithms. | Primarily emphasizes deep learning and strategies that can be used to support generative AI. |
Problem-Solving Emphasis | Addresses a broad spectrum of solutions that often involve statistics and probability. | Primarily used to solve extremely complex problems that have a large number of dependencies. |
Types of Computational Intelligence
Computational intelligence frameworks are inspired by biological systems and evolutionary processes. Over the past decade, researchers have made great strides in using mathematical and programming principles to mimic natural problem-solving processes.
Important frameworks within the field of computational intelligence include:
Neural Networks
Neural networks are inspired by the structure and function of the human brain. They consist of interconnected nodes (neurons) that are able to process and learn from data. Deep learning, a subset of neural networks, has been particularly influential in recent years, leading to significant advancements in new types of deep learning models for image recognition and natural language processing (NLP).
Fuzzy Logic
Fuzzy logic is a powerful tool for being able to still use algorithms when traditional binary logic may not capture the nuances and uncertainties of input data. In this context, the label “fuzzy” means “vague or imprecise.” Fuzzy logic handles uncertainty by allowing variables to have degrees of truth between 0 and 1.
At the conclusion of mathematical operations, fuzzy algorithmic outputs are then “defuzzified” to obtain precise (crisp) numerical results.
Evolutionary Algorithms
Evolutionary algorithms are a type of optimization algorithm inspired by the process of natural selection. The process begins by randomly or heuristically creating a population of potential solutions referred to as candidates. A selection process determines which candidates should be chosen to become parents for the next generation, and the algorithm continues to iterate through generations until a termination condition is met.
Common termination conditions include reaching a maximum number of generations, achieving a satisfactory solution, or running for a specified amount of time.
Swarm Intelligence
Swarm intelligence is inspired by the collective behavior of some types of animals like ants, bees, and birds. This type of connective intelligence simulates the cooperation and interaction observed in natural swarms and adjusts outcomes based on the collective knowledge of the group.
Swarm intelligence is particularly useful for optimizing tasks in dynamic environments when there are lots of interdependent variables.
Cognitive AI
Cognitive AI systems are designed to mimic the human brain’s reasoning processes and understand context. One of the key features of cognitive AI systems is their ability to quickly identify patterns and relationships in extremely large datasets and use what they learn to solve new types of problems in real time.
Developmental AI
Developmental AI systems draw inspiration from learning processes observed in child development. The goal of this type of computational AI is to create systems that can demonstrate a certain degree of domain independence.
Essentially, this type of computational AI starts with a small amount of knowledge and experience and uses interactions and feedback to acquire new skills and knowledge over time.
Computational Intelligence Society
The goal of the Electrical and Electronics Engineers (IEEE) Computational Intelligence Society is to focus attention on the computational and theoretical aspects of mimicking nature to solve problems.
According to the IEEE website, the Computational Intelligence Society (CIS) was officially founded in 2004, and CIS core technologies include neural, fuzzy, and evolutionary computation, as well as hybrid intelligent systems that combine these and other related paradigms.