Basic AI Concepts and Terminology
Vikram Elango, Vivek Gangasani, Shreyas Subramanian · 2025
This chapter examines fundamental concepts in artificial intelligence (AI), tracing its evolution from early symbolic approaches to modern machine learning and deep learning. It begins by defining AI as systems that mimic human intelligence, highlighting key milestones such as the 2024 Nobel Prize in Physics for contributions to machine learning. The discussion covers early AI methods, the shift toward data-driven learning, and the transformative impact of deep learning through neural networks. Distinctions between AI, machine learning, and deep learning are clarified, along with core components like models, algorithms, and data types. The chapter explores various learning paradigms—supervised, unsupervised, reinforcement, and self-supervised—each suited to specific problem domains. Supervised learning, for instance, is illustrated with fraud detection systems, while unsupervised learning is demonstrated in anomaly detection and customer segmentation. Reinforcement learning is exemplified by chess-playing AI, and self-supervised learning is highlighted in speech recognition. The chapter also delves into deep learning architectures, such as CNNs for images and RNNs for sequential data, and the revolutionary impact of transformers on natural language processing.