Understanding AI

Artificial Intelligence (AI) encompasses technologies designed to simulate human thought, behavior, including thinking, speaking, feeling, and reasoning. AI applies computers to areas requiring knowledge, perception, reasoning, understanding, and cognitive ability. For computers to achieve these capabilities, they must understand common sense and facts, manipulate qualitative data, handle exceptions, understand relationships between facts, interact with humans, and adapt to new situations based on prior learning. Unlike information systems that focus on data handling, AI generates and displays knowledge and facts to support decision-making.

Robots are a successful application of AI, with 'soft robots' made from flexible materials emerging as a cost-effective alternative to conventional ones. Expert systems mimic human expertise in specific fields to solve problems in well-defined areas. They consist of programs that replicate human thought behavior in specific problem-solving domains. While decision support systems generate information using data models, expert systems also work with heuristic data. Key components of an expert system include a knowledge acquisition facility, a knowledge base (storing facts, rules, and expectations), a user interface, an explanation facility, and an inference engine that uses forward or backward chaining to derive solutions.

Case-based reasoning (CBR) is a problem-solving technique that matches new problems with previously solved cases. It involves four 'R's: retrieve, reuse, revise, and retain. CBR systems can improve customer service and reduce costs. Intelligent agents, or 'bots,' are software capable of reasoning and following rule-based processes. Sophisticated intelligent agents possess characteristics such as learning from past knowledge, autonomy, collaboration, intuitive interfaces, mobility, and reactivity to environmental stimuli. While most current intelligent agents have limitations, advancements are expected. Applications include web marketing, smart catalogs, shopping and information agents, personal agents, data mining agents, and monitoring and surveillance agents.

Fuzzy logic facilitates a smooth transition between human and computer vocabularies by dealing with variations through a 'degree of membership,' allowing computers to simulate vagueness and uncertainty. This enables computers to reason like humans, using approximations to produce clear answers. Fuzzy logic is applied in search engines, chip design, and database management. Machine learning, or deep learning, involves gaining knowledge through experience. Artificial Neural Networks (ANNs) learn and perform tasks difficult for conventional computers, such as pattern recognition and spam filtering. ANNs, like expert systems, are useful for poorly structured problems with fuzzy or uncertain data. Genetic algorithms (GAs) mimic natural evolution for optimization problems with many input variables, such as jet engine design or portfolio development, and are often combined with neural networks and fuzzy logic systems.

Natural Language Processing (NLP) allows users to communicate with computers using human language, providing natural and easy-to-use question-and-answer settings. Despite the complexity of human language, progress in NLP has led to applications in call routing, trading, and banking. NLP systems perform interfacing (accepting human language input) and knowledge acquisition (summarizing and storing information from text to respond to inquiries). AI technologies, including expert systems, NLP, and ANNs, integrated into decision support systems (DSS) enhance decision quality. This integration improves explanation capabilities, adds learning capabilities, and creates more user-friendly interfaces, especially for non-computer-savvy decision-makers. Virtual assistants like Siri and Alexa exemplify how AI-driven conversational interfaces make DSS more accessible.

Contextual computing describes a pervasive computing environment that senses surroundings and offers recommendations based on user location and relationships. The core principle is that computers can sense and react to their environment, similar to how human brains understand and interpret stimuli, enhancing decision-making based on accumulated experience.